IIOLOGY UBRARY )rnla Institute of Technology 14 - DAYS STUDIES OF INHERITANCE IN GUINEA-PIGS AND RATS BY W. E. CASTLE AND SEWALL WRIGHT PUBLISHED BY THE CARNEGIE INSTITUTION or WASHINGTON WASHINGTON, 1916 CARNEGIE INSTITUTION OF WASHINGTON, PUBLICATION No. 241 PAPER No. 26 OF THE STATION FOR EXPERIMENTAL EVOLUTION AT COLD SPRING HARBOR, NEW YORK FROM THE LABORATORY OF GENETICS OF THE BUSSEY INSTITUTION Copies of this Book were first issued SEP 20 1916 PRESS OF GIBSON BROTHERS, INC. WASHINGTON CONTENTS. PART I. AN EXPEDITION TO THE HOME OF THE GUINEA-PIG AND SOME BREEDING EXPERIMENTS WITH MATERIAL THERE OBTAINED. BY W. E. CASTLE. PAGE. Introduction 3 Some observations on guinea-pigs in Peru 5 Hybridization experiments with Cavia cutleri 8 Life history of C. cutleri 8 Crosses of C. cutleri males with guinea-pig females 13 Color inheritance among the F 2 hybrids 13 (a) Cross 9 albino (race B) X d" C. cutleri 13 (6) Cross 9 albino (race C) X d" C. cutleri 14 (c) Cross 9 brown-eyed cream (race C) X d" C. cutleri 16 (d) Results from (6) and (c) combined 16 (e) Intensity and dilution among the hybrids 17 (/) Significance of the results observed 18 Hybridization experiments with a race of feral guinea-pigs from lea, Peru 20 Origin and characteristics of the lea race 20 Crosses between the lea race and guinea-pigs of race C 23 The F 2 generation 24 Summary on the lea race 29 Hybridization experiments with a domesticated guinea-pig from Arequipa 31 d 1 1002 and his FI offspring 31 F, offspring of d" 1002 33 Back-cross and other offspring of d" 1 1002 35 Miscellaneous matings of the descendants of d" 1002 36 Summary on the Arequipa domesticated race 41 Size inheritance in guinea-pig crosses 42 Previous work on size inheritance 42 Weights and growth curves of C. cutleri, of various guinea-pig races, and of their hybrids 43 Skeletal measurements of C. cutleri, of various races of guinea-pigs, and of their hybrids 47 The C. cutleri hybrids 48 Hybrids of the Arequipa d" 1002 52 The lea hybrids 53 Theoretical explanations of size inheritance and of "blending inheritance" in general 54 PART II. AN INTENSIVE STUDY OF THE INHERITANCE OF COLOR AND OF OTHER COAT CHARACTERS IN GUINEA-PIGS, WITH ESPECIAL REFERENCE TO GRADED VARIATIONS. BY SEW ALL WRIGHT. PAGE. Color and its inheritance in guinea-pigs 59 Skin, fur, and eye colors of guinea-pigs 59 Color of Cavia cutleri 59 Melanin pigment 59 Primary classification of fur colors 59 Yellow group of colors 60 Dark group of colors 60 Skin colors 61 Eye colors 62 Definitions of fur colors by Ridgway's charts 62 Definitions of eye colors 63 Heredity of fur and eye color 63 Color factors of guinea-pigs 63 Classification of color factors . . 63 IV CONTENTS. Color and its inheritance in guinea-pigs Continued. PAGE. Heredity of fur and eye color Continued. Color vs. white 64 Intensity of general color development 64 Dark vs. yellow color 65 Variations of dark color 66 Table of factor combinations 66 Hereditary factors and the physiology of pigment 67 Discussion of experiments 74 Material 74 Systematic position 74 Description of stocks 74 Problems 77 Inheritance of dilution 77 The red-eye factor 77 Dilution 79 The dilution factor 80 Inheritance of minor variations in intensity 85 Methods and accuracy of grading 85 Variations in intense guinea-pigs and albinos 85 Multiple allelomorphs 86 The relations of imperfect dominance, stock, and age to grades of intensity . . 87 Variations of yellow 89 Variations of sepia 92 Variations of eye color 93 Summary " 93 Inheritance of variations in the agouti pattern 94 Previous work 95 The inheritance of the agouti of C. rufescens 96 Minor variations 99 The inheritance of the agouti of C. cutleri 100 Inheritance of rough fur 100 Classification 102 Previous work 102 Material 103 Problems 104 Inheritance of rough as opposed to smooth 105 Inheritance of major variations 106 Possibilities of linkage among rough and color factors 113 Summary of rough tables 115 Minor variations 116 Roughness of series II 117 Summary 118 General conclusion 119 Experimental data 121 Explanation of tables 62 to 137 121 PART III. FURTHER STUDIES OF PIEBALD RATS AND SELECTION, WITH OBSERVATIONS ON GAMETIC COUPLING. BY W. E. CASTLE. PAGE. The progeny of hooded rats twice crossed with wild rats 163-168 A second report on mass selection of the hooded pattern of rats 168-172 Further observations on the mutant series 173-174 Gametic coupling in yellow rats 175-180 Tables 181-187 BIBLIOGRAPHY 188-190 EXPLANATION OP PLATES. . . ... 191-192 PART I AN EXPEDITION TO THE HOME OF THE GUINEA-PIG AND SOME BREEDING EXPERIMENTS WITH MATERIAL THERE OBTAINED BY W. E. CASTLE INTRODUCTION. For several years I have been engaged in studies of heredity in guinea-pigs. In the course of these studies all the common varieties of guinea-pigs have been investigated by the method of experimental breeding and something has been learned concerning their inter- relationships and probable mode of origin. The actual origin of most of these varieties is, however, unknown, as is true also concerning most varieties of domesticated animals. One or two varieties have, however, been made synthetically in the laboratory and it is conceivable that, if we had the original wild stock to work with, from which the domesticated guinea-pig has arisen, some or all of the existing varieties might be synthesized anew and perhaps still others might be obtained, and that in this way something might be learned of the method by which new varieties arise. From considerations such as these I have for several years been seeking to obtain living specimens of the wild species which most closely resemble guinea-pigs. In 1903 I received from Campinas, Brazil, 3 wild-caught individuals referred at the tune to the species Cavia aperea, but since found to agree better with the description of C. rufescens. From two of these animals young were obtained, and crosses, the results of which have been described in detail by Dr. Detlefsen (1914), were made with domesticated guinea- pigs. It may be noted that all male FI hybrids were sterile, but that the FI females were fertile, and that upon repeated crossing of these with male guinea-pigs, a race of fertile hybrids was at last obtained, these being, in the language of breeders, about guinea-pig, | rufescens. From this result it seems doubtful whether C. rufescens has any close genetic relationship to the domesticated guinea-pig, although by hybridization it has been found possible to produce races (f or more guinea-pig) which have derived certain characters from a rufescens ancestor. Cavia aperea from Argentina has been crossed with the guinea-pig by Nehring (1893, 1894) in Berlin, with the production of fully fertile hybrids. This result indicates a closer relationship with the guinea-pig than C. rufescens manifests. Darwin (1876), however, did not regard aperea as the ancestor of the guinea-pig, because he found it to be infested with a different species of louse. I have not myself been able as yet to obtain specimens of C. aperea. Nehring (1889) has argued with much plausibility that Cavia cutleri of Peru is more probably the ancestor of the guinea-pig, for (1) it agrees closely with the guinea-pig in cranial characters and it occurs in a region where guinea-pigs have been for a long time kept in domestication, as is shown by the occurrence of mummified guinea-pigs which had been buried with the dead. Natu- rally I formed a strong desire to secure living specimens of C. cutleri for 3 4 INTRODUCTION. experimental study, but for several years I was unable to do so. Through correspondence with Professor S. I. Bailey, who was at the tune director of the Harvard Astronomical Observatory at Arequipa, Peru, I ascertained that a wild species of cavy occurred in that locality. Professor Bailey kindly captured some of the cavies and attempted repeatedly to forward them to me, but without success. The steam- ship companies refused to accept them for transportation on the ground that they might lead to detention or quarantining of their vessels, since all rodents were suspected of being carriers of bubonic plague. After several years of waiting and fruitless negotiation with every chance traveler to Peru with whom I came in contact, I resolved to go to Peru myself and get the desired specimens. Through a grant made by the Carnegie Institution of Washington I was enabled, in the fall of 1911, to carry this resolution into effect. The Carnegie Institution of Washington and the Bussey Institution have together provided means for carrying out the breeding experi- ments described in this paper. I wish to express my gratitude to both institutions and to thank the director and other officers of the Harvard College Observatory for hospitality and generous assistance given me at the Arequipa station. I am indebted also to Professor C. J. Brues for kindly bringing me a stock of guinea-pigs obtained by him near Lima, Peru, in 1912. SOME OBSERVATIONS ON GUINEA-PIGS IN PERU. On a midsummer day in December 1911 I arrived as a guest at the Harvard College Observatory in Arequipa, Peru, where I went in search of guinea-pigs, wild and domesticated, to be used in breeding experiments. The day after my arrival at the observatory I walked a short dis- tance up the highway through a group of adobe cabins, straw-thatched and without chimney or windows, and with a single door. On looking in at the open door of one of the cabins, I was pleased to see a domesti- cated guinea-pig of the common spotted black-and-white sort familiar to lovers of pet-stock throughout the world. In other near-by cabins I found considerable numbers of guinea-pigs were kept, in one as many as 40. They were fed on fresh-cut alfalfa or the green leaves of maize, receiving apparently no other food and no water. At the back or sides of the cabin was a sort of shelf or bench of stone used as a seat or couch, underneath which the guinea-pigs had their home. Their escape through the open door was prevented by a high lintel of stone, perhaps 15 inches (38 cm.) high, over which one has to step in entering. In these cabins were seen most of the common color varieties of guinea- pigs known to us, agouti, black, yellow, and white (albino). None of the colored individuals which I saw was self-colored; all were spotted with white or with yellow or in both ways. The same predilection for spotting is seen in the other important native domesticated animal, the llama. I saw no llamas except such as were spotted; some were black spotted with white, but the majority were of a soft shade of buff or fawn spotted with white. The common spotted condition of our guinea-pigs is undoubtedly one of long standing; indeed it would seem that the Peruvian natives breed no other variety except such as are either white spotted or all white. The unspotted or " self-colored " varieties now kept by fanciers in Europe and America have probably been produced by selection from stock originally spotted. This is indicated by the great difficulty in securing a self-colored race entirely free from spotted individuals. Most self-colored races, even when bred for many generations from self-colored ancestors exclusively, will pro- duce an occasional individual bearing a few hairs or a patch of hairs of some other color, or of white. Among the guinea-pigs kept by the natives near Arequipa, I observed an occasional animal having a rough or resetted coat. This variety is known to fanciers in Europe and the United States under the name Abyssinian. (See Castle, 1905.) It is said, on the authority of Geoffrey Saint-Hilaire, to have been introduced from Peru into Europe about the year 1872 in a rough-coated, long-haired individual received at the Jardin d'Acclimatation, Paris. In conformity with this account 5 6 INHERITANCE IN GUINEA-PIGS. it may be said that the rough-coated long-haired variety has ever since its introduction been called by fanciers " Peruvian." I saw no long- haired individuals, either rough-coated or smooth, among the guinea- pigs kept by the natives at Arequipa, and the short-haired rough-coated ones observed had imperfectly developed rosettes, much inferior to the best standard-bred resetted Abyssinians of fanciers in Europe and the United States. For this reason I infer that no particular attention was given to this character in the breeding of the guinea-pigs which I saw, though this may very likely have been done in other parts of the country. But the unit-character variation which is responsible for the resetted condition of the coat in Abyssinian guinea-pigs was plainly represented in the stocks kept by the natives in Arequipa and needed only selection to bring it up to the standards of fanciers. Eight independent mendelizing unit-character variations had been recognized as affecting the coat characters of guinea-pigs up to this time. Six of these were represented among the four or five dozen guinea-pigs which I actually saw in the cabins of natives, the other two unit characters being (1) the long-haired variation which, as already noted, is said to have been brought originally from Peru to Europe; and (2) the brown variation which first came to the notice of fanciers in England about 1900 and was certainly in existence before that time in the United States, as I can state from personal knowledge. It is uncertain whether or not this last variation had already occurred in Peru and was thence transferred to Europe, but it is certain that all the other 7 had done so, and it is very probable that this also originated in Peru. Further, a ninth wholly independent unit-character variation (presently to be described, viz, the pink-eyed variation) has made its appearance in stocks of domesticated guinea-pigs obtained by me at Arequipa in 1911 and by my colleague, Professor C. T. Brues, at Luna, in 1912. So it is clear that this variation also is widely disseminated among domesticated guinea-pigs kept by the natives in Peru and which have never been in the hands of European fanciers at all. It can be stated, therefore, with probable correctness, that the guinea- pig has undergone in domestication more extensive variation in color and coat characters than any other mammal, and that this variation has occurred almost if not quite exclusively under the tutelage of the natives of Peru. This conclusion points either to a great antiquity of the guinea-pig as a domesticated animal or to more rapid evolution by unit character variation than by other natural processes. That the natives do give careful attention to the selection of animals for breeding is shown by the following incident : In the cabin near the observatory, where I first saw guinea-pigs in Peru, and where I ulti- mately secured two pairs of animals, one of which I brought back with me, I observed a very large individual which I desired to purchase, and though other individuals were offered me at a very reasonable price, GUINEA-PIGS IN PERU. 7 this particular one could not be had because, I was assured, he was the "padre" (sire) of the entire family. Size seemed to be the point especially emphasized in the breeding of guinea-pigs hi this cabin, as would naturally be the case when the animals formed the meat-supply of the family, as they do now among the native poor of Peru and doubt- less have done since ancient times. But the chief object of my journey to Peru was the study not of the domesticated guinea-pigs of the country, but of their wild progenitors. Accordingly special efforts were made to secure specimens of the wild cavy, which Professor Bailey had found to be abundant hi the locality. Once or twice, when riding along a road between irrigated fields, I had seen a cavy scurry to cover in a pile of rocks; further, I had observed droppings of the animals in the rocky wall of a cattle corral in an alfalfa field. But how to capture the animals alive was a problem which baffled immediate solution. It seemed likely that the natives would know better how to go about this than I did. Accordingly word was passed around among the near-by villages that a good price would be paid at the observatory for wi!4 cavies, either alive or dead. Within a few hours boys began to arrive with the coveted specimens and for the next week I was kept busy preparing skins and saving bones of the animals which were received dead, or making cages and caring for such as arrived alive. In this way 11 cavies (all I could hope to transport safely) and about a dozen skins were soon secured, and preparations were made for the return journey. In due time the journey was accomplished, and with such success that three new races of guinea-pigs were added to our experimental stocks, viz, (1) a wild species, the probable ancestor of the domesticated guinea-pig, identified as Cavia cutleri Bennett; (2) a feral race from lea, probably identical with that described by Von Tschudi; (3) domesticated guinea-pigs, such as are at present kept by the natives of Peru. 8 INHERITANCE IN GUINEA-PIGS. HYBRIDIZATION EXPERIMENTS WITH CAVIA CUTLERI. LIFE HISTORY OF CAVIA CUTLERI. The primary object of my journey to Peru was to secure representa- tives of the wild species of cavy, Cavia cutleri Bennett, known to exist there. Four pairs of these animals captured at Arequipa were suc- cessfully installed in cages at the Bussey Institution in January 1913. One of the males soon died without leaving descendants; the other 7 animals (4 females and 3 males) produced offspring in captivity, which have continued to breed succesfully, though the stock has at times been seriously reduced by disease in cold weather. Three generations of descendants have been reared from the original stock of 7 animals. Together they number 100 individuals, of which 47 are males and 53 females. All are very uniform in color, size, general appearance, and behavior. Their color is a dull leaden gray-brown, well adapted to escape notice amid the arid surroundings of their native habitat. The fur is agouti- ticked and the belly light, but the yellow of the ticking and belly is so pale as to resemble a dirty white or very light cream shade. The color is much paler than that of the Brazilian species, Cavia rufescens, studied by Detlefsen. The fur is also finer and softer, in which respect it resembles the guinea-pig. The size of C. cutleri is about the same as that of C. rufescens, and between one-third and one-half that of the guinea-pig. The maximum weight of an adult male is about 525 grams ; that of a domesticated male guinea-pig obtained in Arequipa (of 1002) is nearly three times this amount. In wildness Cavia cutleri is very much like C. rufescens. The animals live contentedly in small cages, 2 feet 6 inches square, but invariably retreat under their box or conceal themselves in the hay if anyone approaches. The extreme savageness toward each other of individuals of Cavia cutleri makes it difficult to rear large numbers of them in captivity. It is seldom possible to keep more than a single pair in a cage together for any length of tune. Two adult males will not five together peace- ably under any circumstances, and if two females are placed together in a cage with one male persecution of one female by the other usually follows. Even when the young are allowed to grow up in the same cage with their parents, family dissensions are likely to arise as soon as the young become mature. The period of gestation (minimum interval between litters) averages 3 or 4 days shorter than in guinea-pigs, being 60 to 70 days, and the number of young to a litter varies from 1 to 4. Fifty-three litters born in captivity include exactly 100 young, an average of 1.89 young to a litter. The size of litter occurring most frequently is 2, which has been recorded CAVIA CUTLERI. 9 TABLE 1. Number and size of litters produced by each mother, Cavia cutleri. Mother and date of her birth. Date of litter. Size of litter. Mother's age at birth of young. Days since last litter. 9 2 (caught wild) ; born March 1911 (?) Mar. 5, 1913 2 months 24 9 3 (caught wild) ; born May 1911 (?) June 28, 1913 Aug. 29, 1913 Nov. 4, 1913 May 29, 1912 2 2 2 3 27 29 31 12 62 67 9 5 (caught wild) ; born Jan. 1910 (?) Oct. 3, 1912 Dec. 26, 1912 July 5, 1913 Dec. 15, 1913 July 12, 1912 3 1 3 1 3 17 20 26 32 18 9 6 (caught wild) ; born Jan. 1910 (?) Sept. 12, 1912 Nov. 15, 1912 Jan. 22, 1913 Sept. 6, 1912 3 3 2 3 20 22 24 20 62 64 68 9 11; May 29, 1912 Sept. 26, 1912 1 4 9 15; July 12, 1912 Dec. 10, 1912 2 5 926; Sept. 6, 1912 ,. Feb. 17, 1913 June 30, 1913 Oct. 1, 1913 Aug. 15, 1914 Dec. 2, 1914 July 5, 1912 1 3 2 2 1 1 7 12 15 25 29 10 69 927; Sept. 6, 1912 Sept. 4, 1912 Nov. 4, 1912 Apr. 25, 1913 2 2 1 12 14 7 61 61 9 36; Oct. 3, 1912 June 25, 1913 Aug. 25, 1913 June 17, 1913 1 1 1 9 11 8 61 61 942; Nov. 15, 1912 Oct. 16, 1913 Aug. 3, 1914 Nov. 2, 1914 July 26, 1913 4 2 2 2 12 22 25 8 965; Jan. 22, 1913 Sept. 20, 1913 Aug. 25, 1914 Nov. 2, 1914 July 26, 1913 2 2 2 2 10 21 24 6 56 69 966; Jan. 22, 1913 Nov. 1, 1913 Dec. 27, 1913 June 25, 1913 2 3 2 9 11 5 56 9 79; March 5, 1913 Aug. 29, 1913 June 28, 1913 2 1 7 3|t 65 9 118; June 28, 1913 Sept. 4, 1913 Nov. 4, 1913 2 1 6 4 68 9 124; June 30, 1913 Jan. 5, 1914 Nov. 1 1913 1 3 6 4 62 9129; July 5, 1913 Dec. 27, 1913 2 6 9 184; Sept. 4, 1913 Mar. 1, 1914 Mar. 15, 1914 1 1 8 6 64 9224; Oct. 16, 1913 May 18, 1914 July 20, 1914 July 30, 1913 1 2 1 8 10 9 64 63 9 241 ; Nov. 4, 1913 Dec. 8, 1913 Jan. 12, 1915 1 2 14 14 10 INHERITANCE IN GUINEA-PIGS. 24 times; litters of 1 have been recorded 18 times, litters of 3, 10 times, and a litter of 4 once. Factors which influence size of litter are evi- dently age and state of nourishment of the mother. Table 2 shows the relation of age of mother to size of litter. Very young mothers (age 4 months or less) have only 1 young at a birth. The females become sexually mature at a very early age, as do female guinea-pigs. Well- nourished females may breed at 2 months of age, when they are less than half-grown, full growth not being attained until they are 12 or 13 months old. Females over 4 months but under 12 months of age produce usually 1 or 2 young at a birth, rarely 3; those which are 1 or 2 years old produce the maximum number of young, usually 2 or 3, rarely 1 or 4. After the age of 2 years the number of young again TABLE 2. Relation between age of mother and size of litter, Cavia cuileri. Age of mother in months. Size of litters and number of each size. Age of mother in months. Size of litters and number of each size. 1 in litter. 2 in litter. 3 in litter. 4 in litter. 1 in litter. 2in litter. 3 in litter. 4 in litter. 4 3 2 3 1 1 1 2 1 1 12 to 15 1 3 1 2 6 2 2 2 3 1 1 5 16 to 19 6 1 3 3 2 1 1 20 to 23 1 7 24 to 27 8 28 to 31 1 1 9 32 10 Total litters. . . 11 18 24 10 1 decreases to 1 or 2. The oldest female known to have borne young (one of the original stock) had at the time been in captivity over 2 years and her estimated age was 32 months. None of the females born in captivity has given birth to young at a more advanced age than 29 months. Our records accordingly indicate that females rarely breed after they have attained the age of 1\ years. The duration of the breeding period in the case of males is more extended. It is prob- able that males do not attain sexual maturity quite so early as females, for females may breed when less than 2 months old, but we have no evidence that males can breed before they are 3 months old. 1 But the capacity to breed once attained continues indefinitely. One male (cT4) caught wild in December 1911 and estimated then to have been 6 months old is still siring young, more than 3 years after his capture, being, it is estimated, nearly 4 years old. Females are capable of breeding again immediately after the birth of a litter, but if they do so the number of young at the next birth is *Mr. Wright has called my attention to a record from his experiments which shows that a male guinea-pig containing a slight infusion of rufescens blood must have been sexually mature at 2J months of age. This is the only record known to me of a guinea-pig male breeding when less than 3 months old. CAVIA CUTLERI. 11 apt to be less, or the young will be born smaller and less fully developed (with smaller bodies and shorter hair), and the period of gestation will be shortened, even to 56 days in extreme cases, the normal period being, as in the guinea-pig, between 60 and 70 days. (See table 4.) If the mother is well nourished and has not borne a litter recently, she is more likely to have a large litter of young. The largest litter recorded (4) was borne by a female 1 year old, which had previously had only 1 young, born 4 months earlier. The recorded date of the birth of each litter of young is given in table 1, together with the interval in days between suc- cessive litters by the same mother, except in cases where the interval is obviously greater than the ordinary period of gestation, and it is to TABLE 3. Relation of size of litter and number of litters to time of year. Size of litters and number of each size. Total litters. Total young. 1 in litter. 2 in Utter. 3 in litter. 4 in litter. January 1 1 2 1 1 3 2 1 2 1 3 1 3 1 2 5 5 1 4 1 4 7 February March April May June Total born in first 6 months 1 9 6 1 15 22 July 2 1 1 3 5 3 1 5 2 3 2 1 2 1 1 8 6 6 3 8 7 17 11 13 9 17 11 August September October November 1 4 December Total born in sec- ond 6 months 2 . . 9 19 9 1 38 78 1 Average, 1 .47 young per litter. 2 Average, 2.05 young per litter. be supposed that the mother did not breed again immediately. The variation in these day intervals between litters is shown in table 4, from which it appears that the gestation period ordinarily continues from 61 to 69 days, with 63.3 days as an average. However, the periods as recorded can not be relied upon as accurate, except within limits of about 2 days, for the cages were not inspected daily, but only once or twice a week, and when young more than 24 hours old were found in a cage, the estimated age of the young may differ from the true age by 1 or 2 days. Young less than a day old are readily recog- nized as such by the condition of the umbilical cord. The 4 original wild-caught females have a somewhat better record of productiveness than their descendants reared in captivity, which indi- 12 INHERITANCE IN GUINEA-PIGS. cates that laboratory conditions of close captivity are not as favorable for full growth and vigor as the freer life and better air of the original habitat. The 4 wild-caught females produced 33 young in 14 litters, an average of 2.36 young to a litter. Their daughters or granddaughters, reared in captivity, when of like age, have produced 27 young in 13 litters, an average of 2.07 young to a litter. Too much emphasis must not be laid, however, on this difference, because productiveness depends largely on food, care, and weather con- ditions, and it is not certain that these were equally favorable for the original females and for their descendants, respectively. Table 1 shows for each mother how many litters of young she has borne, at what age she bore them, and how many young were contained in each litter. In the case of the 4 females caught wild, the age given for the mother is of course not known; the age recorded is an estimate based on the size of the mother when captured. Table 3 shows in what month each litter of young was born and what its size was. This table brings out rather strikingly the effect of the seasons and consequent character of food available upon the size and number of the litters. TABLE 4. Variation in period of gestation (interval since previous litter) in Cavia cutteri. Average, 63.8 days. Days Days between Cases. between Cases. litters. litters. 56 2 65 1 61 4 67 1 62 3 68 2 63 1 69 2 64 3 In the 6 months from July to December inclusive, litters were born which were conceived under summer conditions, with an abundance of green food available. It will be observed that in this half of the year the litters are numerous (38) and large (average 2.05 young to a litter). The young born in the 6 months January to June inclusive were con- ceived under winter conditions, when the mothers were subsisting largely on a diet of dried or concentrated foods, with a limited amount of green food available. In this half of the year the litters are less numerous (15) and smaller (average size 1.47 young). Temperature probably does not directly affect the result, as the animals were kept in a heated house, but purity of the air may possibly do so, as the house is much better ventilated in the warmer months. But food is probably the most important factor, as the condition of the animals changes promptly with change of food, even when other condi- tions show no change. CROSSES OF CAVIA CTJTLERI. 13 CROSSES OF CAVIA CUTLERI MALES WITH GUINEA-PIG FEMALES. Crosses have been made only between male Cavia cutleri and female guinea-pigs. The reciprocal cross was not undertaken, because the number of cutleri females on hand at any one time has been insufficient and because it seemed probable that a cross with the much larger guinea-pig would be fatal to the cutleri females because of the probable large size of the hybrid offspring. Two races of guinea-pigs were employed in the crosses, these being the purest races available, the genetic properties of which had been long and thoroughly tested. The race most extensively used may be called race C. It consisted of "brown-eyed cream" individuals or of albinos borne by brown-eyed cream parents. The results of crosses of colored and albino individuals of race C will be described separately. The other race may be called race B. It consisted of intensely black-pigmented individuals or of albinos produced by such black individuals. The results of crosses with the two sorts of individuals will be described separately. A cutleri male bred in captivity (cf 78) was mated with black females of race B, and produced 9 FI young, all colored like C. cutleri, but darker, the ticking of the fur being brick red or yellow instead of creamy white as in cutleri. Albino females of race B were mated with the same cutleri male (c?78) or with cfH4, another cutleri male reared in captivity, or else with cT4 or c?8, which were original cutleri males caught wild. Such matings produced 39 FI young, all with the same (golden agouti) type of coloration as the young produced by the black mothers. Females of race C were mated only with the two wild-caught cutleri males (cf4 and cf8). The cream-colored mothers of race C produced 34 young, all golden agouti in color like the young derived from race B crosses, but much lighter. They were, however, darker in color than C. cutleri, the agouti ticking being yellow or reddish, not creamy white as in cutleri. (See plate 3.) Albino females of race C produced by the cutleri males 14 young, indistinguishable in appearance from the young produced by their cream-colored sisters. The FI hybrids, whose total number was 96, were all vigorous and large, their adult size nearly or quite equaling that of guinea-pigs. They grew with great rapidity and have proved fully fertile inter se. In wildness and ferocity they are intermediate between the parent races. COLOR INHERITANCE AMONG THE F 2 HYBRIDS. (a) CROSS 9 ALBINO (RACE B) X rf CUTLERI. By breeding inter se certain of the F! hybrids, from the cross 9 albino, race B, X cf cutleri there has been produced a second (or F 2 ) generation of hybrids, which number 75 individuals. As regards color, disregarding minor differences of intensity of pigmentation, these hybrids fall into three classes: golden agouti, black, and albino. Of the 14 INHERITANCE IN GUINEA-PIGS. agoutis there are 43, of the blacks 15, and of the albinos 17. The numerical relations of the classes suggest a dihybrid Mendelian ratio of 9:3:4, which is in entire agreement with existing knowledge of color inheritance in guinea-pigs (Castle, 1905; Sollas, 1909). C. cutleri is evidently homozygous for all Mendelian color factors, since it breeds very true to color. Albino guinea-pigs from a black race are known to possess two independent recessive modifications from this condition, lacking both the agouti factor and the so-called color factor. As regards these factors, then, the wild race, cutleri, forms gametes AC, the albino forms gametes ac, and the F! hybrids form gametes of the four types AC, Ac, aC, and ac. From recombination of such gametes should arise in F 2 zygotes as in table 5. TABLE 5. 1 AACC 2 AaCC 2 AACc 4 AaCc 9 agouti 1 aaCC 2 aaCc 3 black 1 AAcc 2 Aacc 3 albino 1 albino The several kinds of albinos being similar in appearance, the expected result is 9 agouti, 3 black, 4 albino. The agreement with this expecta- tion is fairly close (see table 6). TABLE 6. Agouti. Black. Albino. Observed 43 15 17 Expected 42.19 14.06 18.75 (b) CROSS 9 ALBINO (RACE C) X C 1C CO OS CO t^ OS 1C "M O IN O 1C t^ rH (N CO * * IN O rH t^ 1C O CO iC CO CO CO CO J^ O X *C CO 1C X CM (N 1C CO CO 6'99-9'99 ^ 99~0 ' 99 O X * CO rH (N (N CO t'* CO CO ^ CO d IN 00 rH t^ CO IN 1C X t^ >* o x r~ rH S" _^ a 03 I s 1 1 1 03 6 '99-9 '99 fr'99-0'99 rH CO CO f*9-0'W i ( I"" iC IN rH CO 6 ' 89-9 ' 89 rH rH CO rH CO Tfl rH f 89-0 ' 89 INrHrH IN CO IN 1C : ** : 6 ' 9-9 ZQ rH rH . (N -* CO 1C f 29-0 S9 : * M : : <* : : : ; ; ~ ; 6 '19-9 '19 I> IN CO 1C rH Oi CM fr'T9-0'I9 TCIO ; 'tf CO rH CO CM rH rH rH 6 '09-9 '09 :* : t-t- - CO i-H CO CM rH rH rH CO r09-0'09 rH rH OS IN IN 1C CM : ^ 6 ' 69-9 ' 69 CO (N OS CO l-H Tfl ** CM IN CO rH * 69-0 69 TtH T* l> rH rH :** IN rH rH : ** 6 '89-9 '89 CO IN CO <*> CO CO fr'89-0'89 CO * CO rH Tf Tf rH rH rH 6\i9-9'Z,9 rH CO CO (N CM CM ; - 1 ^'^9-0^9 IN rH rH (N 1-1 CM (N CM rH 6 '99-9 '99 CO . . . CO ' J ~ t fr'99-0'99 rH - C- co co co co cocococo coco coco cococo cococo '|t?^O T OX^CO t^COO^t* Xit t^CO Xt^^J 1 OXt^ 6-Sfr-S'S* : : - : r^'S* 6-i^-s-i* ~ fl*~OO *OiO>O OOO O lO CO CO CM t^ 00 IN COCO OOO l^OOO rHCOOO (N (N O 10 00 OS * CO O rH O OS O CO b- b- ~" SSSco 1 88 22 cSS 'S SS5 833. 8333 33 $$ 3$ (M CM * CO 1WI W Oi CO CO I s * C*^ Ci ^ CO ^H *O ^ 00 t^ ^* O t^ t* ^ C^ W CO >iO i^ W W ^* W IO ^ ^ ^ Classes (in millimeters) and frequencies. VLWL* : : : : : : : : : : - : : : : j^J 6-9^5-9* CM rH rH ... . rH *'9f-0'W rH rH CM rH 6-s*-s-s* CM (N 00 rH rH *-s*-o-9* rH rH rH O rH r- * rH IN 6-ff-S'W rH Tt< ^ CO ^O rH rH CO rw-o-w - ! COCO- ^"3 OCO rHr- ( ... e-CT-s-e* rH CO rH CO >O rH CO rH O r- H rH CM rH *'8f-0-e* rHM< : ^HCMrH CO^ b- COr- < CM rH CM 6-s*-s-z* rH CM CO (NT}<(N CO-* 00 CMC" 1 S CO -CM *-sFM)-gt 10 Tfl | O CO CO OS rn : : 6-I*-9-I* >O CM CO OS CD CO rjt r)< CM rH fW-O-T* CO CM CO CD Tfl C rH rH rH rH 6-onrw CM (N (N CM rH IO rH rH rH rH *-0*-0'0* CO CO r- * (N IN rH rH ... 6'68-S'68 eo . mm mm c pqpq pq pq & & a 11 11 22 22 " nn ^ h XX XX 2 XX XX .... > 0-0- O 1 O" "CW* 3 * 3 *C W -^ -^ 2 2 1 i H OOO '^ CP O O ^ ^N ^4 ^^ H ~- o "*~" O C8 oooo l b r b r b r b 001- F b r b oc )O OO s fe fe fc > o o o J 0} O O 5 O O O 3 03 03 03 4 t+ mm (X XX 3 c3 08 03 JO O O 1 HH rH rH ^ PR PH P^PH *o r b r b r b 52 INHERITANCE IN GUINEA-PIGS. result from the cross, but (2) that a physiological growth stimulus (not hereditary) results in F : from the fact that the zygotes produced are formed by the union of gametes from very dissimilar races, and (3) that the increased F! vigor is largely, but not entirely, lost in F 2 . No evi- dence is found that it persists in full force in any F 2 zygote (with one possible exception) , since the upper half of the range of the F! zygotes is almost completely wanting in F 2 , while the absence of any appreciable increase of variability in F 2 shows that any increased vigor due to the cross which persists into F 2 persists also very generally among the zygotes of that generation, so that practically all are changed in the same sense and in like amount; otherwise increased variability must result, irrespective of whether size inheritance occurs other than by complete blending. HYBRIDS OF THE AREQUIPA d* 1002. Crosses between the Arequipa cf 1002 and females similar to those of race B, but averaging a little larger, have yielded an extensive and vigorous race for the study of size inheritance. Among the animals of this crossed race the mortality has been comparatively small, so that good numbers are available. The male 1002, sole male ancestor of this race, is still alive, so that his bone measurements are not available; but a female, 1001, secured in the same cabin in Arequipa in 1911, lived until fully grown and her bones are available for comparison. Further, a son of cTl002 and 9 1001 lived until fully grown and his bones also are available. From the measurements of these two and a comparison of the empirical ratio of female to male measurements in the other races studied, it is possible to arrive at estimates of the racial size of the Arequipa stock which it is believed are fairly reliable. These are given in table 32, where it is further assumed that the racial size of the animals mated with cf 1 1002 was substantially that of race B, measurements of the latter being given for comparison. But whether these assumptions are sound or not does not affect the validity of the observations on the F! and F 2 hybrids from this cross, which are valuable as regards their interrelations, for the numbers of adult individuals are considerable (43 F! and 77 F 2 animals) and the mortality among them is small. F! in this experiment (tables 29 to 31, rows 9-12) regularly exceeds the assumed mid-parental measurement, as in the crosses previously considered; F 2 is in all cases close to the mid-parental, being slightly greater in three cases and slightly less in three cases. As regards the relative variability of the two generations, the standard deviations indicate that the F 2 females (as compared with those of FJ are con- siderably more variable in skull length (though scarcely more so than race B) and are slightly more variable in skull width and femur length. The male F 2 hybrids differ very little in variability from the Fj hybrids, the standard deviations being slightly greater in skull measurements but less in femur length. SIZE. 53 THE ICA HYBRIDS. Crosses with the lea race were made principally by a male of race C whose measurements are known and which slightly exceed the averages for race B males; but a few crosses with the lea race were also made with race B females (mated with lea males). The measurements given in table 32 for the mates of the lea race are a mean between the meas- urements of races B and C. The standard deviation of the mixed parents should of course exceed that of race B alone, which should increase the variability of F-i and F 2 , but should not alter the relative variability of these two, since the race B and race C hybrids were bred TABLE 32. Statistical constants derived from tables 23 to 25. Races. No. of indi- vid- uals. Skull-length. Skull-width. Femur-length. Mean. Standard deviation. Mean. Standard deviation. Mean. Standard deviation. 9 cutleri 10 28 24 33 7 63 26 24 51.55 58.14 57.70 54.35 52.91 60.35 61.20 57.26 61.10 58.14 61.92 60.44 63.11 60.35 64.40 61.86 57.45 59.00 60.20 58.84 58.10 62.00 62.20 62.17 13.50 19.75 16.85 17.20 9.45 15.05 12.15 20.00 30.84 34.68 35.24 33.26 31.63 36.33 37.79 35.24 39.70 34.68 38.37 37.35 41.20 36.33 39.40 38.87 36.28 35.00 37.13 35.63 38.00 38.00 39.26 38.73 9.35 10.56 11.60 11.45 6.80 11.90 11.70 12.05 38.45 41.16 42.63 40.38 38.77 42.39 43.57 41.32 45.50 41.16 44.07 42.95 46.50 42.39 45.16 43.15 42.64 42.00 43.49 42.06 42.90 42.50 44.63 43.44 12.05 12.50 10.35 15.60 8.20 10.70 9.80 14.20 9 race B 9 Fi, cutleri X B or C ... 9 F 2 , cutleri X B or C ... cf cutleri cf race B cf Fi, cutleri X B or C ... cf F 2 , cutleri X B or C ... 9 1001 Arequipa 9 race B 28 18 41 63 27 56 8 19.75 12.10 20.65 15.05 15.15 17.75 10.00 10.00 11.20 22.65 23.55 24.55 10.56 9.40 14.05 11.90 11.95 11.40 9.80 6.30 8.70 16.40 12.10 11.80 12.50 13.35 15.30 10.70 10.05 11.80 10.75 9.05 10.85 16.10 17.60 17.80 9 Fi, Arequipa X race B . 9 F2, Arequipa X race B . cf Arequipa (estimated) . . cf race B . . . cf FI, Arequipa X race B . cf F 2 , Arequipa X race B . 9 lea 9 races B and C 9 Fi, lea X race B or C . . . 9 F 2 , lea X race B or C . . . cf lea 7 14 10 cf races B and C . . cf Fi, lea X race B or C . . . cf F 2 , lea X race B or C . . . 8 17 entirely distinct from each other and are only tabulated together to secure greater numbers. Again in this cross we are confronted with the same phenomena as regards the skeletal measurements: (1) in F! a substantial increase in size (least in skull length of FI male hybrids) ; (2) while in F 2 a return is made toward the mid-parental (mean of the races crossed), in 4 of the 6 measurements it is closely approximated; (3) the standard deviation meanwhile alters little, not enough to have significance ; the lea crosses are of particular interest because the races mated are of nearly the same size. The phenomenon of increased size in F! followed by a prompt loss of the increase in F 2 is here observed 54 INHERITANCE IN GUINEA-PIGS. exactly as in the crosses between races of widely different and heritably different sizes, but without indication in either case that the size inheri- tance is other than a simple and permanent blend. THEORETICAL EXPLANATIONS OF SIZE INHERITANCE AND OF BLENDING INHERITANCE IN GENERAL. We conclude therefore that, so far as present knowledge goes, the statement made in 1909 that size inheritance is blending and does not mendelize still holds. This does not preclude the possibility that in special cases mendelizing factors may exist which affect size. For example, in man brachydactyly is due to such a factor, a simple Mendelian dominant, as was first shown by Farabee (1905), and has been confirmed by Drinkwater in the case of three separate English families. This character involves a shortening of the skeleton generally, but of the digits in particular. It is transmitted only through affected individuals, the normal offspring of affected individuals producing only normals. Professor James Wilson has stated that the Dexter-Kerry cattle of Ireland differ from ordinary Kerry cattle by a similar men- delizing factor. If one were to restrict his study of size inheritance to cases such as these, he would reach the conclusion that size inheri- tance in general is Mendelian, a wholly mistaken idea. (See Castle, 1914.) Such cases among animals are distinctly rare. Among culti- vated plants they seem to be somewhat commoner, so that many of the inherited size differences studied by botanists involve such factors. One of the commonest of these is involved in the difference between normal (tall) and dwarf habit of growth, a case demonstrated by Mendel for peas in his original experiments ; but it is more than doubtful whether Mendelian factors produce the differences in height observed among different races of tall or of dwarf peas respectively. The same is true concerning differences in size or shape of seeds and fruits, as described by Emerson and Gross. It seems almost certain that Mendelian factors are involved in many of the cases studied, but associated with other factors not Mendelian, possibly merely physio- logical, which render the results extremely complex and the variation seemingly continuous in character. To have shown that size inheri- tance is occasionally affected by Mendelian factors is not by any means to have demonstrated that all size inheritance is due to Mendelian factors. The physiological increase of size due to the crossing of unre- lated races is a fact of far greater economic importance to the animal breeder than the existence of any Mendelian factor affecting size that has thus far been demonstrated. The question may be raised, how are we to account for the increased variability of F 2 as compared with F 1; if this is not due to segregation and recombination of multiple factors, as assumed under the Nilsson- Ehle principle. (1) This would be sufficiently accounted for in the SIZE. 55 case under discussion by an unequal persistence, among the F 2 zygotes, of the increased growth stimulus observed in FI and due evidently to the act of crossing, not to inheritance. (2) Increased variability in F 2 would also result if a blending occurs in P\, which is imperfect, so that the gametes formed by the F : individuals are not all the exact mean of the parental gametes, but fluctuate around that mean. What may we imagine the germinal basis of a blending character to be? Perhaps some substance or ferment which varies in amount, larger amounts producing larger results. If a 5 per cent solution of cane sugar were poured into the same dish with a 10 per cent solution and then sam- ples were dipped from this before the two solutions had been thoroughly stirred together, it might very well happen that the samples would not be of uniform strength. Any other result would be surprising. A char- acter genuinely blending in heredity might be expected to behave in this same way, the quantitatively different conditions found in parent races not blending perfectly in a single generation of association together in an FI zygote, which therefore would produce gametes less uniform in character than those of the respective inbred parent races. The multiple factor interpretation of size inheritance, besides being superfluous, meets with this serious logical difficulty: If we suppose the difference between two races to depend upon a certain number of independent factors whose action is cumulative, then a less difference must be due to fewer factors, and the fewer factors concerned hi a cross, the more obvious is the segregation. But we do not find it easy to detect segregation when races are crossed which differ little in size; the general result is the same as when races are crossed which differ widely from each other. It is difficult to detect any evidences of segregation unless the parent races differ widely from each other, under which condition, if multiple factors are involved, complete segregation should occur least often. On the whole, the hypothesis of quantitative variations in a blending character presents fewer difficulties as an explanation of size inheritance than the hypothesis of multiple unvarying segregating factors. It is to be preferred on the ground of simplicity alone, but it also accords better with the results obtained in other fields. Jennings now finds, contrary to his earlier observations on paramecium, which Calkins and Gregory were unable to confirm, that size is a character varying even in asexual reproduction, within what would be a "pure line" if the theory of factorial constancy were true. My own observations of rats and other rodents (Castle, 1915) may be cited to show that even single Mendelian unit characters are quantitatively variable. If this is so, the hypothesis of multiple factors as a general explanation of variability is quite unnecessary and so should be discarded. PART II AN INTENSIVE STUDY OF THE INHERITANCE OF COLOR AND OF OTHER COAT CHARACTERS IN GUINEA- PIGS, WITH ESPECIAL REFERENCE TO GRADED VARIATIONS. BY SEWALL WRIGHT, S. D. COLOR AND ITS INHERITANCE IN GUINEA-PIGS. The experiments described in the following paper were carried on at the Bussey Institution of Harvard University, between September 1912 and August 1915, under the direction of Professor W. E. Castle. A large number of stocks of guinea-pigs and wild cavies, containing an extensive assortment of variations, were available throughout the experiments, and furnished excellent material for studies on inheritance. The writer wishes here to express his gratitude for the privilege of using freely this material and for the constant encouragement and assistance which Professor Castle has given. SKIN, FUR, AND EYE COLORS OF GUINEA-PIGS. COLOR OF CAVIA CUTLERI. The fur color of Cavia cutleri, the probable ancestor of the guinea-pig, is of the agouti type found hi most wild rodents, as well as in many other wild mammals. (See plate 3.) The back and sides are slaty black, ticked with yellow (more accurately, cinnamon buff). An isolated hair is of a dull slate color at the base, becoming blacker toward the tip. Near the tip there is a yellow band some 2 or 3 mm. long. The extreme tip for 1 to 2 mm. is black. The belly is cream-colored (more accurately cartridge buff) and is sharply separated from the ticked sides. An isolated hair is pale neutral gray throughout its basal half and cream- colored in the remaining portion. Cavia rufescens of Brazil has a similar ticked coat, but differs hi showing less ticking on the back and sides and often in having a ticked belly not sharply separated from the sides. The general appearance is darker. Tame guinea-pigs show a great variety of colors and color patterns and also deviations from the dark skin and black eye color of the wild species. MELANIN PIGMENT. The coat colors of mammals are largely due to granular pigments of a kind known chemically as melanin. The pigment in the hair is found principally in the walls of air-spaces in the medulla, but to some extent in the cortex, as described by Bateson (1903) in mice. Melanin pigments are also found in the skin (principally in the epidermis) and in the iris and retina of the eye. A deficiency of pigment hi the retina is revealed by a red reflection through the pupil. PRIMARY CLASSIFICATION OF FUR COLORS. Three qualitatively distinct melanin pigments are generally recog- nized in mammals, viz, black, brown, and yellow (Bateson, 1903) . There are reasons, however, for regarding black and brown as more closely related to each other than either is to yellow. Black and brown granules are acted upon similarly by most hereditary factors which act 59 60 INHERITANCE IN GUINEA-PIGS. on either. Yellow pigment, on the other hand, is acted upon very dif- ferently from black and brown by many factors. Accordingly it will be convenient to use a term to include both black and brown pigments, as dark pigments. The fur colors fall naturally into two groups, the dark and yellow colors, characterized by the predominant presence of dark and yellow pigments respectively. YELLOW GROUP OF COLORS. In the yellow group of colors the one of highest intensity is a rich yellow-orange, which matches quite well with ochraceous tawny in Ridgway's color charts (1912). There are all gradations from this ochraceous tawny through cinnamon buff and cartridge buff to white. In this paper it will be more convenient to use the conventional names, red, yellow, and cream, for these grades. In grading the guinea-pigs, three samples of hair have been used as standards of grades called red , yellow 3 , and creamg, respectively. White is considered to be cream 8 . All of the yellow colors in guinea-pigs fall into this series, as far as known. In mice, however, Little (1911) has shown that two dilution series between red and white can be distinguished. There is a series from red to cream resembling in appearance (though not geneti- cally) the guinea-pig series. Another series (the " dilute" reds, yellows, and creams) has a peculiar streaky appearance. The physical relation between these two series is probably similar to that between the sepia and blue types of dilution among the dark colors, which is discussed below. DARK GROUP OF COLORS. Among the dark colors there are at least three distinct series : (1) There is the series of neutral grays, passing from black to white. Such colors are shown by the blue rabbits, blue mice, and maltese cats. There are no tame guinea-pigs known whose colors fall distinctly into this series ; but the dull black of the wild Cavia cutleri, especially on the belly, is a neutral gray quite free from any brown. Examination of the hair of the blue rabbit under the microscope shows dense black pigment masses alternating with colorless spaces, a condition described by Miss Sollas (1909) in the hair of the blue mouse and apparently comparable to the clumped condition of the black pigment in the feathers of blue pigeons, described by Cole (1914). (2) There is a series of grades from black through dull brown and tow-color to white. This series is shown by dilute black guinea-pigs. The various shades of human hair, from black through brown to tow- color, match samples from this guinea-pig series very closely. The increase in quantity of pigment in this series in passing up from the lower grades is accompanied by a change in quality. Yellowish-brown pigment gives way to black. Dilution of this sort is produced inde- COLOR. 61 pendently by different factors, the combination of which gives doubly dilute colors, which may still be classed in the same series. Dilute guinea-pigs of this series have been called blue in the literature, but the name is as inappropriate as it would be applied to human brown hair, and, moreover, tends to confusion with the very distinct type of dilution of the blue rabbit. In this paper the colors of this series will be called sepia. Grades of dilution have been represented by numbers, as in the yellow series. White is considered as grade 16. Grading has been done by comparison with standard samples of hair, the colors of which are defined in terms of Ridgway's colors at the end of this section. (3) The most intense grade of this series is a rich dark brown, such as is found hi chocolate guinea-pigs, mice, and rabbits and in liver- colored dogs. This color is not very different from sepia4, but is some- what warmer and less dull. As noted by Miss Durham hi the case of brown mice, there seems to be a complete absence of black granules, but a large quantity of brown granules. No intergrades between this brown and black are known. There are dilute browns, each corre- sponding closely to a color hi the sepia series. They are often difficult to distinguish from grades of sepia in isolated samples of hah*. On the animals, however, the browns seem conspicuously richer than the sepias. There are, further, correlated differences in skin and eye color which are even more conspicuous. Most guinea-pig colors can be matched fairly well in the sepia, brown, or yellow series, but one other class of variations must be noted. The animals have been graded by the color near the tip of the hair, but while in some blacks, sepias, browns, and yellows the hah* is nearly uniform, hi most cases the base is much duller than the tip. This gives a somewhat streaky effect to the fur. In the case of dull blacks of this kind, the color at the base of the fur is usually between a neutral slaty black and a dark sepia. SKIN COLORS. The color of the skin usually corresponds roughly to the color of the hair which comes from it. Where the fur is thick there is very little pigment in the skin, while exposed places (as ears and feet) are often very strongly pigmented. Where the fur is yellow the skin in exposed places shows an orange- yellow color, usually with considerable admixture of black. On most of the body the skin is white, with occasional orange-yellow spots. The dilution of fur color is accompanied by dilution of the skin color. Where the fur is black the exposed parts of the skin are very black, while the rest of the skin is dull black. Where the fur is of the sepia series the color of the ears and feet depends much on the genetic factors responsible for the dilution of the black. In the sepias of the albino 62 INHERITANCE IN GUINEA-PIGS. series the ears and feet are quite black, often with intense black blotches. Even in albinos, where the fur is nearly pure white, the ears and feet may be black. In the pink-eyed sepias, on the other hand, there is very little pigment anywhere in the skin. Brown fur goes with a uniform brown color of ears and feet very different from the dull black of sepias of corresponding intensity of fur color. Dilution in the skin accompanies dilution in the fur. The different skin colors are very conspicuous in animals with spotted fur. In these it is easy to find places where the skin spots do not correspond exactly to the fur spots. White fur may arise from colored skin and yellow fur from black skin, but the reverse cases do not seem to occur. EYE COLORS. The iris and retina usually contain black and brown pigment. Where there is reduction of pigment in the iris, the pigment tends to disappear first next to the pupil, leaving a dark outside ring. Decreas- ing grades of retinal pigment are most easily recognized by the apparent color of the pupil. In black eyes the pupil appears black. Occasion- ally a red reflection can be obtained in strong light. In brown eyes a dark-red reflection is easily obtained by holding the guinea-pig away from the light. In the red eye the pupil looks red most of the time and the inner ring of the iris often transmits red light. A pink eye has a transparent iris and a pink reflection is visible through both iris and pupil in all lights. The following summary shows the color terms to be used in this paper, with their nearest equivalent on Ridgway's color charts (1912). The numbers 15'i, etc., refer to the position in Ridgway's system. For purposes of convenience in defining the color factors, white is included as a member of each color series as well as in a class by itself. In some cases white may be shown to represent extreme dilution of a particular color; in other cases it stands in no relation to particular colors. DEFINITION OF FUR COLORS BY RIDGWAY'S CHARTS. 1 . Pigment absent because of factors not belonging to a dilution series. White. 2. Pigment present, or absent only because of factors demonstrably belonging to a dilution series. a. Yellow group. Redo =15% ochraceous tawny. Yellow 3 = 16"6, redder than cinnamon buff, 17"6. Cream 6 = 19"/, cartridge buff. White. b. Dark group. (1) Black. Slaty black = dark neutral gray. Blue = neutral gray. White. COLOR. 63 b. Darfc'group Continued. (2) Black. Sepia 3 = 16"'n, warmer and darker than clove brown, 17'" m. Sepia 6 =16"% warmer and lighter than clove brown, 17"' m. Sepia 9 = 17'"% hair brown, slightly purer, however. Sepia l2 = 17""&, light drab, somewhat purer. Sepia 1 5 = 17 // "/, pale drab gray, somewhat purer. White. 8^ (3) Brown = 15" m, bister, 15"m, but somewhat wanner and duller. Brown 3 = 15"% between army brown, 13"% and buffy brown, 17"'i. Brown 6 = 17'"6, somewhat duller than avellaneous, 17"'&. Brown 9 =17""/? White. DEFINITIONS OF EYE COLORS. (l)*Black: black iris and pupil. Dark red : black iris, dark-red pupil in favorable lights. Red : partially transparent iris, red pupil in most lights. Pink: transparent iris, pink reflection through both iris and pupil. (2) Brown: brown iris, dark-red pupil. Brown-red: partially transparent brown iris, red pupil. Pink : as above. HEREDITY OF FUR AND EYE COLOR. COLOR FACTORS OF GUINEA-PIGS. Considerable work has been done on the inheritance of color varia- tions in guinea-pigs. The numerous colors which have been listed and several patterns in which these colors may be arranged have been found to be due in the main to relatively few hereditary factors. Some of these factors determine effects which are very easily defined. Thus, any guinea-pig which is homozygous for factor C a is an albino with pink eyes and white fur, regardless of the presence of any combination of other known factors. On the other hand, certain factors determine nothing except in combination with other factors. Factor E may be present in guinea-pigs of any known color variety whatever. It can only be said that its presence is a necessary condition for the develop- ment of more than a trace of dark pigmentation in the fur. The color which results from a given combination of factors can be made clear most easily by classifying the factors into a series of groups. The following classification is based upon the factors in the rodents which have been most studied, viz, guinea-pigs, mice, rats, and rabbits. CLASSIFICATION OF COLOR FACTORS. 1. Factors which affect the distribution and intensity of color largely irre- spective of the kind of color. A. Factors which govern the distribution of color as opposed to no color (white) in patterns in the fur, in individual hairs, and in the eyes. B. Factors which govern the intensity of general color development within colored areas of fur and eyes. 2. Factors which govern the differentiation between yellow and dark colors in colored areas of the fur. 3. Factors which determine the kind of dark color in the areas with dark pigmentation in fur and eyes, without influence on yellow areas. 64 INHERITANCE IN GUINEA-PIGS. COLOR VS. WHITE (1 A). Probably dilution of the type of the blue and dilute yellow mice and rabbits and maltese cats belongs here, rather than in IB, since the effect seems to be due to the distribution of pigment within the indi- vidual hairs rather than to any effect on the actual pigment granules. Most of the factors which belong hi this class, however, are those which determine patterns of white as opposed to areas which are colored under most combinations of other factors. In this class are such fac- tors as on the one hand determine a self-colored coat, and on the other black-eyed whites, as in mice; white patterns, as in hooded rats, Dutch and English rabbits; or scattered white hairs, as hi silvered guinea- pigs. In cases where several independently inherited white patterns have arisen it is evident that there can be no single factor which alone determines self. The "self " allelomorphs of the white-pattern factors can merely be defined as conditions for self. Where more than one white-pattern factor is present in an animal, combination patterns are produced. Clear-cut Mendelian factors which belong to this group are known in mice, rats, and rabbits, but none have been isolated in guinea-pigs, although irregular blotching and silvering with white are common. The symbol S will be used to represent an assemblage of unanalyzed factors. Stc, an assemblage of unanalyzed factors which determine white spotting. INTENSITY OF GENERAL COLOR DEVELOPMENT (1 B). In this group fall albinism and its variations. These factors affect all color, but not wholly irrespective of the kind of color. There are several peculiarities which are discussed more fully in a later section (page 70). The most important is the fact that the level of intensity of the color factor at which yellow can develop at all is higher than the threshold for black or brown. This does not affect the differentiation of the fur into yellow and dark pigmentation areas by factors of group 2, but involves the result that with certain albino series factors, yellow areas appear white, while dark areas are quite strongly colored. Indeed, in albinism itself, dark pigmentation areas can often be distinguished from yellow areas by a slight sootiness in the former, absent in the latter. C. Determines the highest intensity of color of skin, fur, and eye which is to be found with a given array of other factors; dominant over Ca, CV, and C a , where distinguishable in its effects. In the following table, and hi the similar tables under Cd, CV, and C a , are given the ranges of intensity in the yellow, black, and brown series to which these colors develop when the factor under consideration is present. In the case of black and brown, factor P is assumed to be present. When p is present, black and brown undergo a two-fold dilution. P is also considered present in the case of eye-color. Yellow series redo to yellow2 in guinea-pigs; yellows to creams in Cavia cutleri. Black series blacko to black2. Brown series browno to brown2- Eye color black, brown. COLOR. 65 Cd' Determines an intensity of yellow distinctly lower than does C, an intensity of dark pig- mentation usually, but not always lower than does C, and an intensity of eye color rarely distinguishable from that determined by C. More or less dominant over Cr and C a where distinguishable. (Wright, 1915.) Yellow series yellow2 to creamy. Black series blacko to sepiay. Brown series browno (?) to brown? . Eye color black, brown. C r . Determines the complete absence of yellow, an intensity of dark pigmentation indis- tinguishable from that determined by Cd and an intensity of eye color lower than that determined by C or Cd- More or less dominant over C a where distinguish- able. (Castle, 1914o; Wright, 1915.) Yellow series white. Black series blacko to sepias . Brown series browno (?) to browny. Eye color red, brown-red. C . Determines an absence of pigment, complete with yellow, not quite complete with dark pigments of the fur and skin, but complete in the eyes. (Castle and Allen, 1903; Castle, 1905; Sollas, 1909; Detlefsen, 1914; Wright, 1915.) Yellow series white. Black series white, dark smudges on nose, ears, and feet. Brown series white, brown smudges on nose, ears, and feet. Eye color pink. DARK VS. YELLOW COLOR (2). Factors of this group affect skin and fur color, but not eye color. In this group come the factors responsible for self yellows, tortoise-shells, and brindles, on the one hand, and self blacks or browns on the other, as contrasted with the ticked or agouti patterns of the wild rodents. Where more than one factor is present which determines a yellow pattern, combination effects are produced, such as in yellow-spotted agoutis among guinea-pigs. The following factors are known in guinea- pigs: E. A condition for more than a trace of dark pigmentation in the fur; determines dark pig- mentation wherever yellow is not determined by other factors; dominant over e, found in the wild species, all agoutis, blacks, browns, etc., but very rarely in self yellows. . Determines the presence of one of the yellow colors in all colored areas of the fur, aside from a slight sootiness; responsible for the yellow in most self yellows, for the white in red-eyed whites, etc. (Castle, 1905, 1907, 1907o; Sollas, 1909; Detlefsen, 1914.) A. Determines the presence of a yellow color in the light-bellied agouti pattern wherever there is dark pigmentation in which the yellow group ticking may show; dominant over A' and a, found in Cavia cutteri and light-bellied agouti guinea-pigs, includ- ing the red-eyed silver agoutis, in which the agouti pattern is in white. A'. Determines the presence of yellow colors in a more restricted agouti pattern than does A, a pattern usually characterized by a ticked belly not sharply distinct from the sides in color; dominant over a, found hi Cavia rufescens and in ticked-bellied agouti hybrids which have rufescens ancestry. (Detlefsen, 1914.) a. Determines the absence of yellow group ticking in hairs of dark pigmentation; found in blacks, browns, etc. (Castle, 1905, 1907, 1907a, 1913; Sollas, 1909; Detlefsen, 1914.) Zy. An assemblage of unanalyzed factors which determine the presence of spots of a yellow color, conditional on factors of group (1) ; found in black and yellow tortoise-sheila, black, yellow, and white tricolors, and in some red-eyed black and white bicolors; probably responsible for an occasional self yellow, though never in the writer's experience. 66 INHERITANCE IN GUINEA-PIGS. VARIATIONS OF DARK COLOR (3) Factors of this group are responsible for browns and pink-eyed sepias, as compared with blacks, hi guinea-pigs; for browns and pink- eyed sepias in mice, and for the new pink-eyed and red-eyed dilute variations hi rats. Where more than one factor of this group or of group IB determines dilution, combination effects are produced. Thus we have very pale sepias resulting from the combined effects of two independent dilution factors (C d C d pp). B. Determines a color of the black-sepia series wherever dark pigmentation develops, including the eyes; has no influence where yellow pigmentation develops; domi- nant over b, present in the wild species and in blacks, sepias, albinos with black points, black-eyed yellows, etc. b. Determines a color of the brown series wherever dark pigmentation develops, including the eyes; has no influence where yellow pigmentation develops; present in browns, brown-eyed yellows, etc. (Castle, 1907o, 1908; Sollas, 1909; Detlefsen, 1914.) P. A condition for intense development of dark pigmentation in the fur and for eye colors more intense than pink; not necessary for intense development of yellow; domi- nant over p. p. Determines a low development of dark colors, i. e., below sepias; has no influence where yellow develops; determines pink eye color. (Castle, 1914o.) TABLE OF FACTOR COMBINATIONS. In determining the color which corresponds to a given array of factors the groups of factors must be considered in the order given. Table 33 gives a list of the color varieties corresponding to the combinations of Mendelian factors. At the top and left of the table are indicated, by symbols, the factors present in each of the varieties named in the body of the table. The color of spots produced by Sw and Sy are given below. Only the varieties marked with an asterisk have not yet been synthesized. These include the pink-eyed yellows and creams and a kind of pink-eyed white which is expected to be indistinguishable from an albino in appearance, though breeding wholly differently. The pink- eyed brown series (bbpp) has not yet been produced and is not included. Some of the varieties have names given by fanciers which have been used in the literature. In this table, however, it seemed best to use a consistent scheme of naming, indicating at once the color and pattern. Agouti is used as the name for a pattern, the banding of hairs of pre- dominantly a dark color with a yellow color. The names preceding agouti give the two colors in each hair. The following table of syno- nyms may be useful: Black-red agouti = golden agouti. Sepia-yellow agouti = yellow agouti. Sepia-cream agouti = silver agouti. Brown-red agouti = cinnamon. Brown-cream agouti = light cinnamon. Sepia = blue. ' ^* Brown = chocolate. Brown eye= brown eye (Castle), ruby eye (Sollas). COLOR. TABLE 33. 67 Factors Fur. Eve present. EA (agouti light-belly). EA' (agouti ticked-belly). Eaa. ee (A, A' or aa) . B PC Black-red agouti Black Red Black. CdCd Dark sepia-yellow agouti Dark sepia .... Yellow Do. CdCr Dark sepia-cream agouti . Do Cream Do. CnCa Light sepia-cream agouti Light sepia .... ... Do Do. CrCr Dark sepia-white agouti Dark sepia White (light Red. Light sepia-white agouti Light sepia points) . Do Do CaCft ^^hite (dark points) White (dark Do Pink Bpp C Pale sepia-red agouti points) . Pale sepia Red Pink. CdCd Very pale sepia-yellow agouti Very pale sepia . *Yellow Do. CrlCr Very pale sepia-cream agouti. . . Do ... *Cream . . Do CdC a . ... Do Do ...*Do Do. CrCr Very pale sepia-white agouti .Do *White Do .. .Do Do .. .*Do Do. C C White (light points) White (light ...*Do Do. bbP C Brown-red agouti . . points). Brown Red Brown CdCd Medium brown-yellow agouti . Medium brown . Yellow Do. C a C r . . Medium brown-cream agouti . Do Cream Do. CdC a . Light brown-cream agouti .... Light brown. . . . Do Do. CrCr Medium brown-white agouti . . Medium brown . White Brown-red . Light brown-white agouti . . . Light brown . . . Do Do. C 1 (~* White (It. br. points) White (It. br. . Do . Pink. points). Factors present. Sw. Sy. SwSy. Eye. c W^hite spots (clear) . . Red spots . Red and white tri- CdCd Do Yellow spots . . . color. Yellow and white. CdC r . . . Do Cream spots .... Cream and white CdC a . - . .Do Do tricolor. Do C r C r ... . .Do \White spots, Sooty and clear, .Do / often sooty. white spots. (Albino) (Albino) (Albino) HEREDITARY FACTORS AND THE PHYSIOLOGY OF PIGMENT. The definitions which have been given for the hereditary factors are based largely on the colors as seen without a microscope. It would be very desirable, however, to correlate color factors accurately with the variations in quality and quantity of the actual pigments and ultimately with the physiology and chemistry of pigment formation. Considerable progress has been made in recent years in the study of the chemistry of melanin pigments. The melanins are amorphous granular pigments found throughout the animal kingdom. A large number of researches have established the fact that substances which closely resemble the natural melanins can be produced by the action of 68 INHERITANCE IN GUINEA-PIGS. certain oxidizing enzymes on tyrosin and related aromatic compounds. Tyrosin is an important constituent of protein molecules and there is much reason to believe that tyrosin and related substances are the chromogens from which the natural melanins are formed. Tyrosinase, an enzyme, which can oxidize tyrosin to dark substances resembling melanins, has been found very widely among animals, including the skins of mammals, as will be discussed later. There have been many theories on the mode of origin of pigment hi the cells. Early observations indicated that melanin was directly extruded from the nucleus. Recent studies by Hooker (1915) on in vitro cultures of mesenchyme and epithelium of the frog indicate that melanin granules form in the cytoplasm but at the point of known greatest efficiency of the nucleus as an oxidizing agent. Thus, prob- ably chromogen (tyrosin or derivatives) is in the cytoplasm, while oxidizing enzymes are given off by the nucleus. The color white in the fur of mammals is due to the absence of pigment. The theory of a white melanin seems effectively disproved (Gortner, 1910). A priori, the presence or absence of pigment might be conceived as due either to a deficiency of chromogen or of enzyme. In line with the first view, Gortner (1911) found that the pattern in the elytra of potato beetles is due to a deficiency of chromogen. Fur- ther, Cue"not (1903, 1904), hi the first attempt to correlate the facts of Mendelian inheritance with the physiology of pigment, suggested pro- visionally that albinos lack the power of producing chromogen, while the different colors which he demonstrated could be transmitted through albinos depend on specific enzymes. On the other hand, recent observations by Onslow (1915) demonstrate that absence of pigment in widely different cases in mammals depends on enzyme differ- ences. He found peroxidases hi the skins of gray, black, blue, and brown rabbits, which produce a black pigment from tyrosin in the presence of hydrogen peroxide. In the skins of albino rabbits and mice and in the white part of the Dutch pattern in rabbits, all recessive whites, he was unable to demonstrate a peroxidase, although there was nothing present which prevented the oxidation of tyrosin to a black pigment when tyrosinase was added. In the white of the English rabbit, a dominant white, he did find an anti-tyrosinase. Finally, there is strong genetic evidence that albinism in guinea-pigs is not due to absence of chromogen. A diminution in quantity of chromogen should bring about the same diminution hi quantity of all pigments, regardless of quality. But hi red-eyed guinea-pigs (which we may consider as incomplete albinos, as they have an allelomorph of albinism C r ) no yellow develops, leaving white areas where factors of group 2 determine yellow differentiation, but there may be nearly as much black as hi normal guinea-pigs. Indeed, hi the albino guinea- pigs and Himalayan rabbits, there is no yellow, but some black. COLOR. 69 The physical or chemical differences between the pigments of the different fur colors are not wholly clear. According to Onslow (1915) the pigments of black, brown, and yellow rabbits can not be distin- guished, physically or chemically, when isolated. At first sight this seems hardly possible with such apparently different colors. A result thoroughly in line with this view, however, followed the matching of fur colors with Ridgway's charts, much to the writer's surprise at the time. Ridgway distinguishes 72 hues passing from red through orange, yellow, green, blue, and purple, back to red. The yellows, sepias, and browns of guinea-pigs and human brown and red hair all matched colors near hue 17, "orange yellow," in the classification. The differences depended merely on differing amounts of black and white. Bateson (1903), indeed, found that yellow pigment is dissolved from hair by potassium hydroxide very much more rapidly than brown pigment, which dissolved more rapidly than black. This, however, might be due merely to size or density of granules. This apparent qualitative difference in pigments has been attributed to several causes: (1) to variations hi the chromogen acted on by a given enzyme, (2) to interruptions at different stages in the process of oxidation of a given chromogen, (3) to specific enzymes which in each case can only produce a certain result once the action on the chromo- gen is begun. Observations of Onslow indicate that for qualitative differences, as well as for the absence of pigment, enzyme and not chromogen differ- ences are responsible. He could find no peroxidase in self yellows, a recessive variation. There must of course have been some peroxidase at some time to produce pigment at all. Perhaps the apparent absence indicates a very low degree of stability in the yellow-producing enzyme. Again, grays differ from blacks by a dominant factor which causes yellow to appear in ticking over the back but white to appear on the belly. Onslow found a tyrosinase inhibitor in the belly and compared the case with that of the dominant white of the English rabbit. As grays differ from self blacks by only one Mendelian factor, it would seem likely that all of the changes in appearance dorsal yellow tick- ing, ventral white are to be ascribed to one physiological cause. If black is absent from the belly because of an enzyme inhibitor, it would seem likely that black is replaced by yellow in the dorsal ticking by the presence, for a certain period hi the development of a hair, of the same enzyme inhibitor, which, however, is in this case merely an inhibitor of the black-producing reaction, not of the yellow. Reasons for which yellow can appear on the back of rabbits, but not on the belly, when black is inhibited will be discussed later. Thus, a recessive yellow and a dominant yellow-pattern factor are both due to enzyme, not chromo- gen, differences. 70 INHERITANCE IN GUINEA-PIGS. The second hypothesis that yellow, brown, and black are due to interruptions of the normal process of oxidation at different stages is difficult to reconcile satisfactorily with the genetic facts in guinea-pigs. If brown and black pigments pass through a yellow stage, identical with the final stage of the pigment in yellow guinea-pigs, any factor which inhibits the development of yellow must a fortiori inhibit the development of brown and black. We have seen that with factor C r there is complete absence of yellow pigment, but nearly full develop- ment of brown and black. We find nearly the converse of this in the effect of factor p. When factor p is present, the development of brown and black is very greatly reduced without the slightest dilution of yellow. This indicates that neither is yellow a stage in the develop- ment of black nor black a stage in the development of yellow. The most satisfactory hypothesis is the third that there are distinct enzymes which produce yellow and dark pigment. There are a number of curious facts hi connection with the albino series of factors in guinea-pigs which perhaps warrant further specula- tion. As has been mentioned, Onslow has shown that albinism is due to the absence of tyrosinase hi the skin (and presumably the eye). It seems reasonable to suppose that the higher allelomorphs are quantita- tive variations in a factor which determines the power of producing tyrosinase. If this is so, we would expect to find that the different zygotic formulae could be arranged in a linear series with respect to their effects on pigments of all sorts. Following are the series with respect to black pigment of eye and fur, and yellow of the fur. (See plates 1 and 2.) Formula. Black eye. Black fur. Yellow fur. cc Black .... Black Red. CCd ... Do Do Do. CC r .. Do .... Do Do. CC a . . . .Do .... Do Do. CdCd . . .Do .... Dark sepia Yellow. CdC r . ..Do Do Cream. CdC a ... Do Light sepia Do. C r Cr Red Dark sepia . . . White. C r Ca ..Do Light sepia Do. C a C a Pink White (sooty) . . . Do. The yellow series and the less accurately known eye-color series can be arranged in the same sequence. There is the striking difference, however, that the level of no pigment production is much higher in yellow than eye color. The black of the fur agrees with eye color in the level at which pigmentation becomes evident between C a C a and COLOR. 71 C r C a but the sequence can not be made to agree with either the eye- color or yellow series. CaC a is distinctly lighter than C r C r in black fur but distinctly more intense in eye color while, hi yellow fur CaC a is above, C r C r below the threshold of any color. The effects could be explained by a complicated linkage hypothesis. We would need to suppose that there are separate series of allelomorphs acting on yellow, black of fur, and black of eye, respectively, and that C r and C a are complexes identical hi the yellow-dilution factor, C d and C r identical in the black-fur-dilution factor and perhaps C and Cd hi the black-eye- dilution factor. But an hypothesis according to which it is a mere accident that the factors which dilute yellow, black of fur, and black of eye are perfectly linked in inheritance can hardly be taken seriously. Another escape would be to suppose that our four factors, C a , C r , Cd, and C, are, indeed, variations of the same thing but not linear quantita- tive variations. However, it seems most satisfactory to the writer to attempt to explain the results on the basis of four quantitative gradations of one factor, which determines the amount of the basic color-producing enzyme, if it is in any way possible. Let us see what assumptions must be made to do this. First, it will be convenient to assume with Little (1913) that the basic color-producing enzyme (I) acting by itself on chromogen, produces yellow pigment. The addition of a second substance (II) makes it a black-producing enzyme (I-II). We will further assume that I is relatively unstable and must be produced above a certain rate (that determined by C r C r ) in order to reach and oxidize the chromogen in the cytoplasm. United with II it becomes more stable and even produces some effect at the rate of production determined by C a C a . The next assumption is that above the thresh- old for yellow, I-II and the excess of I compete for the chromogen. As a result of partial displacement by the paler color, the intensity of black decreases just above the yellow threshold. CdC a seems paler (and somewhat browner) than C r C r . In the eye, no factor ever brings out a yellow color. There is perhaps never an excess of I here and the intensity of black follows the normal sequence. Summarizing, the hypothesis to which consideration of the physio- logical and genetic seems to lead is as follows : (1) There is a basic color-producing enzyme (I) which acting alone on chromogen produces a diffuse or finely granular pigment which appears yellow. It is relatively unstable. Intensity of production and absence or inhibition in parts of fur and eye are determined by the various factors of group 1 the albino series, " blue "-dilution factors, and recessive and dominant white-pattern factors. (2) There is a second substance (II) which may unite with I to produce a more stable enzyme, which reacts with chromogen to produce a coarsely granular pigment which appears sepia, brown, or black. When II is present, I is stabilized to such an extent that pigment is 72 INHERITANCE IN GUINEA-PIGS. produced at a lower rate of production of I than is the case of I alone. Above the level at which I alone produces yellow, the two kinds of enzymes, yellow- and black-producing, compete with each other for chromogen, producing a mixture of black and yellow, the relative importance depending on the rate at which II is produced. Because of the competition the intensity of black shows two maxima as production increases one just below the yellow threshold and the other at maximal production of I. Intensity of production or inhibi- tion of II in patterns hi the fur are determined by various factors (group 2) which produce self yellow, yellow spotting, agouti, etc. (3) There is a third group of substances which, added to the dark- pigment-producing enzyme (II), affect the intensity of dark color pro- duced but not the power of fixing chromogen in competition with the yellow-producing enzyme. They have no effect on the intensity of yellow. In this group are the brown factors of mice and guinea-pigs, and perhaps rabbits and dogs, the pink-eye factor of rats, mice, and guinea-pigs and the new red-eye factor of rats,^.e., the factors of group 3. While based to a larger extent on the genetic facts in the albino series in guinea-pigs, the hypothesis explains many cases in other mammals in the sense that apparently complex variations are reduced to a single physiological cause. In rabbits, single Mendelian factors produce some rather complex variations. A single factor changes a self black to the gray color with a yellow-ticked back but a pure white belly. Another variation changes a self black to a sooty yellow with a black belly. These varia- tions combined hi one animal give a white-bellied clear yellow. How can each of these apparently complex color changes be determined by a simple physiological change? Let us suppose that in all rabbits I is produced strongly on the back, but so feebly on the belly that it is below the yellow threshold, but not so feebly that black is greatly affected. Let us suppose that II is likewise more strongly produced on back than on belly. A factor which tends to produce an inhibitor of II is added. On the back II (the black-producing enzyme) is inhibited in only a portion of the development of the hair, leaving yellow ticking. On the belly all II is inhibited, leaving white. The result is a gray rabbit. The other factor causes a general slowing up in the production of II. On the back this enables the yellow-pro- ducing enzyme to predominate in competition and sooty yellow results. On the belly below the yellow threshold what little black-producing enzyme does develop has no competition and only black can result. We get a black-bellied sooty yellow. The combination pattern can only be a white-bellied yellow. In many other mammals color phases are found which can be explained as due either to variations in production of II or I. The red phase of the red fox, has a white chest. The level of production of I is below the yellow threshold COLOR. 73 but above the black threshold on the chest. Increase in production of II produces the silver phase, nearly self sepia in color, including the chest. The colors of the varying hare seem to be due to variations in production of I determined by environmental causes. The white winter pelage gives way in blotches to a white-ticked sepia; this gives way to yellow-ticked sepia as the intensity of production of the basic enzyme rises above the yellow threshold and in some varieties the full summer pelage is almost self red. Many other cases could be given in which two color phases of an individual animal or the color patterns of closely allied varieties seem to differ in many respects and yet can be explained on the basis outlined as due to a single physiological change. In the case of very complex color patterns, it is necessary to suppose that the power of producing the hypothetical enzymes I, II, or III may be distributed in quite complex patterns. But the hypothesis often gives a simple explanation for certain peculiarities in a pattern. In the tiger, the stripes on the back are quite intense yellow and black. The yellow stripes grow paler down the sides, becoming white on the lower sides and belly. The black stripes likewise grow lighter down the sides but at the point at which the yellow becomes white, the black stripes suddenly grow more intense, at least in some individuals, to become paler again on the belly. Again, on the legs, which are white on the inside, yellow on the outside, black stripes are visible on the white part, but either disappear completely or leave merely a streak of sooty red on the yellow part. All of this becomes intelligible if we assume that the basic enzyme (I) is produced at decreasing rates from back to belly and from outside to inside of leg, while the black-produc- ing supplement (II) is distributed in vertical stripes (horizontal on the legs). Two parallel stripes give a remarkable reproduction of the variation hi black and yellow in the albino series in guinea-pigs. We have the same change from black and intense yellow to sepia and cream, then to darker sepia and white, and finally light sepia and white, illustrating the different thresholds for the appearance of black and yellow and the reduction in intensity of black above the yellow threshold due to the entrance of competition at this point. DISCUSSION OF EXPERIMENTS. MATERIAL. SYSTEMATIC POSITION. Guinea-pigs belong to the family Caviidse of the hystricomorph division of rodents. There are three living genera of Caviidse: Doli- chotis Desm., which contains the large Patagonian cavies; Hydrochcerus Brisson, to which belongs the capybara; and Cavia Pallas, containing the small cavies. Genus Cavia is divided into two subgenera, Cavia proper and Cerodon F. Cuv., distinguished most conspicuously by the greater complexity of the molars hi the former. Seven living species are listed under Cavia proper by Trouessart (1904) : C. rufescens Lund, a small dark Brazilian cavy with subspecies in Guiana and Argentina. C. fidgida Wagler, a Brazilian cavy probably closely allied to rufescens (Thomas, 1901). C. aperea ErxL, a large pale-colored Brazilian cavy. C. azarce Wagner, a cavy of Paraguay probably closely allied to aperea (Thomas, 1901). C. cutleri Bennett, a small pale-colored cavy of Peru. C. tschudii Fitzinger, a large, richly colored cavy described from lea, Peru. C. porcellus Linn., the tame guinea-pig, much larger than at least rufescens and cutleri. DESCRIPTION OF STOCKS. Four of these species are dealt with hi the experiments to be described, viz, Cavia rufescens, C. cutleri, C. porcellus, and a type which is quite certainly that described as C. tschudii, although it is also quite certain that it is simply feral porcellus. Breeding experiments have been carried on with a fifth species, C. aperea, by Nehring (1894). The C. rufescens stock was derived from 3 individuals received from Mr. Adolph Hempel, of Campinas, Brazil, in 1903. The history of this stock is fully described by Detlefsen (1914). When received by the writer, most of the stock consisted of hybrids containing only from TV * TTT rufescens blood. There were a few \ and | bloods and one f blood, 9 A68, which is still alive (August 1915) at the remarkable age of 8 years 1 month, 1 a good illustration of the vigor of the first-generation hybrids. All of the pure rufescens stock has died out. The rufescens hybrids have been crossed with nearly all of the guinea-pig stocks to be described, and most of the color varieties may be found among them. The ticked-bellied type of agouti has been found only among them and in pure rufescens. C. rufescens was not completely fertile with the guinea-pigs (Detlefsen, 1914). Detlefsen found that while the female hybrids were fertile, all of the male hybrids obtained were sterile. In the | rufescens, derived by crossing the females with guinea-pigs, the males were again all sterile. Not until the | bloods were obtained did October 1915, aged 8 years, 3 months. W. E. C. 74 MATERIAL. 75 a few fertile males appear. The percentage of fertile males gradually increased in later generations. The Cavia cutleri stock was derived from animals captured by Pro- fessor Castle in Peru in 191 1 . Like C. rufescens, these are much smaller than the guinea-pig. All show the agouti pattern. The color is described on page 59. Unlike C. rufescens, C. cutleri breeds freely hi captivity and crosses readily with the guinea-pig. The male and female hybrids are fertile. The lea stock of guinea-pigs was derived from 3 guinea-pigs which were obtained by Castle near lea, Peru, in 1911. They were as large as or larger than average guinea-pigs, and of a rich golden agouti color, very different from C. cutleri. Two independent color variations appeared at once in the pure stock, viz, black (aa) and red-eye (C r C r ). Such variations are very uncommon among wild species of animals; e. g., none has occurred within the pure rufescens or cutleri stocks. Both of these variations are found in domesticated guinea-pigs in Peru (Arequipa stock). From the description of Cavia tschudii, quoted in Waterhouse (1848) under the name C. cutleri Tschudi, it seems clear that our lea stock is the same as the former, which was likewise described from lea. In view, however, of the size, color, and possession of recessive color varieties found among tame guinea-pigs of Peru, there can be little doubt that they are feral porcellus. The Arequipa stock comes from a pair of guinea-pigs brought from Arequipa, Peru, by Castle in 1911. He obtained them from Indians who had them under domestication. Owing to the early death of the only female, no pure stock could be developed, but numerous descen- dants have been derived from the original male 1002, a sepia-cream agouti with white and cream spots, demonstrated to be of constitution EEAaBBPpC d C r , and from a son of the original pair, male 1007, a yellow agouti with white and yellow spots, demonstrated to be of con- stitution EeAaBBPPCjjCd. These were crossed mainly with the 4-toe and BW stocks, which are described below. For a full discussion of the origin and nature of the pure cutkri, lea, and Arequipa stocks, see Parti. The Lima stock comes from 8 guinea-pigs obtained from Indians near Lima, Peru, by Professor Brues in 1913. These guinea-pigs and their descendants have only recently been crossed with other stocks. There have been no agoutis in this stock. The pink-eye and yellow variations, as well as white spotting (but not yellow spotting), have occurred in this stock. A pink-eyed red, the lowest recessive, of this stock is of constitution eeaaBBppCC. There were both rough-furred and smooth- furred individuals in the original stock. The following stocks come from guinea-pigs obtained from fanciers by Professor Castle and have been maintained for several years at the Bussey Institution. 76 INHERITANCE IN GUINEA-PIGS. BB stock. A stock consisting exclusively of very intense blacks. No red or white spotting has been observed among them. Unfortu- nately it is a stock of low fertility, and could not be used much to advantage. BW stock. This stock has for years consisted exclusively of very intense blacks and very sooty albinos. The blacks occasionally show a few red hairs or a small red patch. This has been an extremely useful stock, among other things, furnishing albinos known to be geneti- cally identical with blacks, except for the albino factor. (Race B of Parti.) Four-toe stock. This is a much-inbred stock, practically all the indi- viduals of which show four good toes on the hind feet instead of the normal three. This stock was developed by selection and inbreeding by Professor Castle (Castle, 1906). Most of the individuals are a dull black with dull red blotching and brindling and often with white spots. Albinos appear quite frequently and reds much more rarely. TABLE 34. Genetic formula of stocks. Stock. . Color. Roughness. Mendelian. Unanalyzed. Men- delian. Unana- lyzed. Cavia cutleri Cavia rufescens .... lea E A B P C E A' B P C E A,a B P C,C r E,e A,a B P,p C,Cd,C r E,e a B P,p C E a B P C E a B P C,C a E(e)a B P C,C a E a B,b P C E,e a B P Cd,C a e a b P Cd,C a E,e A,A',aB,b P C,Cd,C r ,C a 2int r S r S r S R,r s R,r (S)s r s r s R,r s R,r S,s 2- .... 2- .... .... 2R .... 2R 2- .... 2- .... .... 2R 2+ .... (2w2y) 2int + 2w2y 2int + 2w Arequipa Lima BB 2int + BW (2y) 2int + 2wSy Sint - 2wSy 4-toe Tricolor Sepia + cream (2w2y) 2int- 2int r s r s R,r S,s Brown-eyed cream . . C. rufescens hybrids . 2w2y 2int In the tricolor stock the fur is typically a patchwork of red, white, and black. Full-roughs, partial-roughs, and smooths occur among them. The writer has used many guinea-pigs of very mongrel ancestry, which, however, owe their partial rough coat to this stock. The sepia-and-cream and brown-eyed cream stocks have been selected for years for extreme dilution. The former stock consists exclusively of sepias, black-eyed yellows and creams, and albinos. The latter consists exclusively of brown-eyed yellows and creams and albinos. (Race C of Part I.) In the tables, these together are called dilute- selection stock. Table 34 shows the Mendelian factors affecting color and roughness of fur which occur in each stock. Unanalyzed hereditary conditions which affect color and roughness are also included, prefixed by the symbol S. Sw and Sy, as has already been stated, mean hereditary INHERITANCE OF DILUTION. 77 white and yellow spotting respectively. 2 int+ and 2 int mean hereditary constitutions which intensify or dilute, respectively, the color associated with a given array of Mendelian factors. 2 + and 2 in the rough column have a similar meaning with respect to the rough character. 2R means the presence of roughness of a different kind from that analyzed. Where a factor occurs only rarely in a stock, it is inclosed in parentheses. PROBLEMS. The inheritance of the discontinuous color variations which are known in guinea-pigs has been solved by previous work. After each factor variation from the wild type (Cavia cutleri) in the definitions of the factors the principal papers on the subject are given. The writer has been concerned mainly with an analysis of inheritance in the contin- uous series of variations by which each of the intense colors red, brown, and black grade into dilute colors and ultimately white. A second group of problems concerns the variations in the amount of yellow ticking in agoutis. The writer has worked with the agouti patterns of Cavia cutleri, C. rufescens hybrids, and tame guinea-pigs. The inheritance of variations in the rough coat occasionally found in guinea-pigs is discussed in a later section. INHERITANCE OF DILUTION. THE RED-EYE FACTOR. The experiments with dilution have become closely associated with experiments with certain imported South American stocks (lea, Are- quipa) which are discussed in detail in Part I. A number of hitherto unknown color varieties appeared in these stocks, the inheritance of which could be explained by assuming the existence of a new allelo- morph of albinism intermediate in effect and dominance between albin- ism and its normal allelomorph. More specifically, this new factor is characterized by the production of red eyes, slight dilution of black in the fur, and complete inhibition of yellow pigment development. The writer has used animals of both the lea and Arequipa stocks in experiments, with results in full agreement with those given in Part I. Crosses 20-1 and 21 to 25 involve red-eye (from lea stock) without also involving dilution. In cross 20-1 a pure lea male, a red-eyed agouti, is crossed with intense guinea-pigs, giving young all intense. This illustrates the dominance of intensity over red-eye. In cross 21 a pure lea intense male crossed with albinos of intense stock gives both intense and red-eye young. The lea male no doubt was heterozygous for red-eye, but the albinos could not possibly trans- mit red-eye, as they come from a stock hi which red-eye has never appeared. This illustrates the apparent reversal of dominance of red- eye whenever albinism is introduced into a cross. A further illus- tration is given in cross 23, in which red-eye by albino of intense 78 INHERITANCE IN GUINEA-PIGS. stock gives red-eyes, but no intense young. In cross 25, red-eyes crossed with albinos from various sources give no intense young, but only red-eyes and albinos. One possible explanation of these results would be the supposition that red-eye becomes dominant over its normal allelomorph in the presence of heterozygous albinism. In this case intense young should appear when such heterozygous red-eyes are crossed together; but, as is shown in cross 24, none such appears. Here red-eyes from cross 21, mated inter se, gave 17 red-eyes, 6 albinos, no intense. Numerous results of this kind have made it clear that intensity can never be recovered in any generation after a cross of red- eye with albino. This means that neither red-eye nor albino can trans- mit the normal allelomorph of the other. Now, the one thing which a recessive variation, of necessity, can not transmit, is its own normal allelomorph. Therefore the normal allelomorphs of red-eye and albino must be identical. This does not yet demonstrate that albinism, red-eye, and intensity form a series of three allelomorphs. There is still the possibility that red-eye and albinism involve the same recessive allelomorph (C a ) of normal color (C), but differ by an independent modifying factor. Symbolically we could suppose albinos to be C a C a rr, red-eyes to be C a C a RR (or C a C a Rr), intense guinea-pigs of ordinary stocks to be CCrr (or CC a rr) , and intense guinea-pigs of lea stock to be CCRR (or CC a RR) . We must suppose the lea stock to be homozygous for the modifier R, to account for the absence of albinos. R must be a unit factor to account for the simple 3 to 1 ratio in cross 24. This hypothesis fits all of the facts given so far. The critical test of its truth is the possibility (as it turns out, impossibility) of producing intense animals (CC a Rr) which will give both red-eyes and albinos when crossed with albinos. If intensity, red-eye, and albinism are triple allelomorphs, it should be impossible to obtain such animals. Crosses 21 and 22 are interesting as furnishing just this test. Cross 21 may be represented symbolically as follows, according to the two hypotheses: Albino (BW) X intense (lea) = 9 intense + 4 red-eye. (1) CaCarr X CCaRR = CC a Rr CaCaRr. (2) CaCa X CCr = CCa CrCa. In either case the Fj. red-eyes crossed inter se should give 3 red-eyes to 1 albino. The result obtained in cross 24 (17 red-eyes to 6 albinos) is in nearly perfect agreement. But the cross of F! intense with albinos gives very different results under the two hypotheses (cross 22) : Albino X intense (Fi) = 16 intense + (0 red-eye) + 25 albinos. (1) CaCarr X CCaRr = CCaRr + CCair + CaCaRr + CaCarr. (2) CaCa X CC a = CC a + CaCa- The complete absence of red-eyes among the 41 young, as well as the excess of albinos where an excess of intense is expected, thoroughly eliminates the first hypothesis. The results agree reasonably well with INHERITANCE OF DILUTION. 79 the hypothesis of triple allelomorphs, which we have found to agree with the results of all the other crosses. The only possibility which has not been eliminated is a linkage so close as to simulate a system of triple allelomorphs. Unless exceptions occur which require it, such an hypothesis need not be considered. Thus the data obtained by the writer are not only in harmony with the theory that albinism, red-eye, and intensity form a series of triple allelomorphs, but can be explained on no other basis, barring the pos- sibility just noted. DILUTION. Such color varieties as agouti, black, brown, yellow, etc., are sharply distinct from each other. They segregate from crosses without pro- ducing intergrades and in unforced agreement with Mendelian expec- tation. In contrast with these discontinuous variations are the con- tinuous variations in the intensity of color of each main color variety. Thus, among the yellows, there are all gradations from a pale cream to an intense red. Among the agoutis, there are the pale silver agoutis, the intense golden agoutis, and the intermediate yellow agoutis. There are all grades of dilute blacks known to the fanciers as blues, for which term, as has been explained, sepia is substituted in this paper. Finally, there are all grades of dilute browns and cinnamons. (See plates 1 and 2.) The existence of these dilute types was noted by Castle (1905) and Sollas (1909), both of whom also recognized that dilution could be transferred from one series to another, e. g., from creams to blacks, giving rise to sepias. They did not, however, suggest any factorial explanation, finding the results of crosses highly irregular. Detlefsen (1914) considered dilution to be recessive, but found the inheritance of dilution very irregular among C. rufescens hybrids. He obtained dilutes in F! after crossing dilute hybrids with a race of guinea-pigs (brindle or 4-toe), among which dilution had never occurred and which therefore should not carry it as a recessive. It may be remarked hi passing that the 4-toe race does contain albinism, which, with present knowledge, satisfactorily accounts for these F! dilutes. Thus the difficulties in the way of an understanding of the heredity of dilution have been due (1) to the intergrading of dilute with intense; (2) to data which seemed to indicate that dilution could be due neither to a recessive nor to a dominant unit factor, without complications. Cross 39 gives many examples in which intense by intense has given very dilute young, which seems to indicate that dilution must be recessive if simple Mendelian at all. On the other hand, such cases as that given by Detlefsen are difficult to interpret on this basis. Further, dilute by dilute has often given young much more intense than either parent. Thus, in cross 42-8, we have two medium sepias producing a black. In cross 37 are many cases in which cream by cream 80 INHERITANCE IN GUINEA-PIGS. has produced yellow. Apparently intense by intense may give any intensity whatever, and almost the same can be said of dilute by dilute. Amid this confusion, however, one cross has been found which con- sistently gives a very definite, although unexpected, result. It is found that a dilute crossed with an albino, even of intense stock, never gives intense young, but only well-defined dilutes and albinos. There are only a few possible ways in which this result can be explained and, from the results of other crosses, all but one of these explanations have been definitely eliminated, namely, that dilution is an allelomorph of albinism. An allelomorph of albinism was already known to be respon- sible for the red-eyed condition in certain South American stocks (Castle, 1914a) . It could now be shown that albinism, red-eye, dilution, and intensity are due to a series of four allelomorphs with dominance in the order of increasing pigmentation. A preliminary account of this demonstration has been given in a previous paper (Wright, 1915). In the present paper the demonstration is given in more detail and further steps are taken in the analysis of the variations. THE DILUTION FACTOR. The dilute varieties have some resemblance to the red-eyed varieties. The fact that red-eye is due to an allelomorph of albinism suggested that dilution might also be due to a member of the same series of alle- lomorphs. A stock was chosen which was known to carry no dilution. This was the BW stock, which for years has consisted exclusively of the most intense blacks and sooty albinos. The following crosses were designed to eliminate the hypothesis of allelomorphism if incorrect : (1) Albinos from intense stock were crossed with dilutes: CaCall X CCii = CCali. (2) Albinos from dilute stock were crossed with blacks of intense stock : CaCaii X CCII = CC a Ii. If intensity and dilution form a pair of allelomorphs (I, i) which segregate independently of the pair color and albinism (C, CJ, as is the case in mice and rabbits, these two crosses must give identical results. In each case, color is introduced by one parent, albinism by the other; intensity by one parent, dilution by the other. In fact, identical results should be obtained regardless of whether dilution is due to a unit factor or to multiple factors, or even whether its inheritance is Mendelian or not, provided only that it is inherited independently of albinism. Crosses 16 and 17 and table 35 give the actual results. All cases are included, which involve an intense stock known to carry no dilution. Among those called dilute below (among the young), none was more intense than sepia 2 or yellow 4 . Among the intense, none was more dilute than a dull black comparable in grade but not in color with sepia 2 , or a red in very few if any cases as dilute as yellow 2 . There was there- fore no difficulty in drawing a natural line between intense and dilute in these crosses. INHERITANCE OF DILUTION. 81 It is evident that the two sets of crosses give consistently different results. This difference demonstrates that dilution does not segregate independently of albinism. An even more striking result follows from a portion of the above data. Fj dilutes, one of whose parents was of intense stock, were back-crossed with albinos of intense stock. They gave 9 dilute, 20 albino young, no intense, although these young were at least three-quarters of intense stock. On the other hand, Fi intense, one of whose parents was an albino of dilute stock, were back-crossed with albinos of dilute stock. They gave 5 intense, 7 albinos, no dilutes, although these young were at least three-quarters of dilute stock. It is clear that the hereditary difference between a dilute and an intense can not be transmitted through an albino. TABLE 35. Intense. Dilute. Red-eyed. White. cf albino (intense stock) X 9 dilute 56 5 39 9 albino (intense stock) X cf dilute 29 10 21 Total 85 15 60 d* albino (dilute stock) X 9 intense (intense stock) . . 9 albino (dilute stock) X c? 1 in tense (intense stock) . . 47 9 10 2 56 12 It was emphasized above that all the intense animals used in cross 17 came from stocks which have never given dilutes. This was necessary because in other crosses (18, 34, 41) intense by albino has given many dilute young. No such precaution was taken with the dilutes used in cross 16. Any available dilutes were used regardless of ancestry. In fact, 1 1 of them, with 38 young, had one or both parents intense. In none of the other crosses in which dilute has been crossed with albino (19, 27, 38, 44) has any intense young appeared. Thus in crosses with albinos an intense may transmit dilution, but a dilute never trans- mits intensity. From these crosses it seems clear that intensity is dom- inant over dilution. Other crosses on the whole bear this out. The apparent exceptions will be ignored for the present but discussed later. We have reached the definite conclusion that dilute by albino can never give intense, regardless of ancestry on either side. Since the only thing which a variety of necessity can not transmit is a dominant allelomorph of its essential factor, it follows that dilution and albinism must have the same dominant allelomorph, which we will call intensity. There are only a few hypotheses which will satisfy this condition. We already know two recessive allelomorphs of intensity, viz, red-eye and albinism. It is conceivable that dilution may be due to the cooperation of an independent factor (or factors) with one or more of the known combinations C r C r , C r C a , and C a C a . If this is not the case, 82 INHERITANCE IN GUINEA-PIGS. dilution must be due to a new allelomorph in the albino series, let us say Cd- A modifying factor which shows partial coupling would give intermediate results. (1) Since dilution and red-eye show considerable resemblance, it would be a plausible hypothesis to assume that they are due to the same allelomorph in the albino series (C r ) but differ by an independent modi- fying factor (D). With this hypothesis, all stocks used (except the lea and Arequipa) must needs be homozygous for the modifier in order that no red-eyes should appear. Dilutes would be C r C r DD or C r C a DD albinos C a C a DD in these stocks. Thus albinos of these stocks should transmit the modifier and in crosses with red-eyes (C r C r dd) should produce dilutes at least in F 2 . But in crosses 23 and 25, red-eyes mated with such albinos have given no dilutes, nor have dilutes appeared in F 2 in cross 24, among 23 young. Thus an albino can not transmit the hereditary difference between a dilute and a red-eye and the hypothesis is untenable. (2) Next to be considered is the hypothesis that there is a modifier which converts into a dilute an annual which would otherwise be an albino. Dilutes of ordinary stock would be C a C a DD or C a C a Dd. In cross 20, dilutes of ordinary stock crossed with a pure lea male No. 724, a homozygous red-eye (C r C r dd), produced 5 dilute young which must be of formula C r C a Dd. This shows that if the hypothesis is to stand at all, it must be extended, so that the factor which converts an albino into a dilute also converts a red-eye into a dilute. The fact that a dilute may transmit red-eye (crosses 19 and 27) is further evi- dence that this extension is necessary. In this form most of the results can be explained satisfactorily. (3) The only other hypothesis which remains is that dilution is due to a new allelomorph in the albino series making a series of four C, Cd, C r , and C a . The results cited above (crosses 20, 19, and 27) make it evident that dilution is dominant over red-eye. The meaning of a series of four allelomorphs can be made clear by considering all of the possible zygotic formulae. Every zygote must have two representatives from the series, but never more than two. Intense guinea-pigs may be homozygous (CC), or carry dilution (CCd), or red-eye (CC r ), or albin- ism (CC a ), but can never transmit more than one of the recessive conditions. Dilutes may be homozygous (CdCd) or carry red-eye (CdC r ) or albinism (CdCJ, never both. Red-eyes may be homozygous (C r C r ) or carry albinism (C r C a ), while albinos can only be homozygous (C a C a ) and can never transmit any of the higher conditions. The critical test between this hypothesis of four allelomorphs and the preceding one (that dilute is a modified red-eye or albino), lies in the possibility or impossibility of producing animals which in crosses with albinos will transmit more than one recessive condition." If an intense animal can be obtained which transmits both dilution and red-eye INHERITANCE OF DILUTION. 83 (CC r Dd) or dilution and albinism (CC a Dd), or if a dilute can be obtained which transmits both red-eye and albinism (C r C a Dd), the hypothesis of modifiers must be adopted. But all attempts to obtain these double heterozygotes have failed. All of the results substantiate the hypothesis of quadruple allelomorphs. Arequipa male No. 1007 was of formula C r C r DD or C d Cd, depending on the hypothesis chosen (see crosses 28 to 34). He was crossed with intense guinea-pigs of BW or 4-toe stock, known to transmit no dilu- tion (CC a dd or CC a ) . The intense young could only be CC r Dd or CCd under the two hypotheses. Five of them were crossed with albinos and gave 13 intense, 20 dilute young, no others (cross 34). Expectation on the hypothesis of a moolifier is 16 intense, 8 dilute, 8 red-eye. On the hypothesis of allelomorphs it is 16 intense to 16 dilute. Both the excess of dilutes and the absence of red-eyes point conclusively to the latter. In cross 18, intense guinea-pigs, each of which had a dilute parent known to transmit albinism and with no lea or Arequipa blood, are crossed with albinos or red-eyes. Under the modifier hypothesis we would expect about half of them to be CC a Dd. Under the allelomorph hypothesis, they should be CCa or CC a . As it turned out, there were 6 which gave only intense and dilute (30 intense, 35 dilute) and 8 which gave no dilute young (57 intense to 61 red-eye or albino). Thus there was no intense which had dilute young and also red-eyes or albinos. This result distinctly favors the hypothesis of allelomorphs. In crosses 19 and 27 dilutes, each from the cross of a red-eye with a stock guinea-pig free from South American ancestry, are crossed with albinos. Under the modifier hypothesis, those which transmit red-eye at all are necessarily C r C a Dd, for they must be C ra C ra D in order to appear dilute; they could get C a , but not C r , from the stock guinea-pig parent, and they would necessarily get d from the red-eye parent. Under the allelomorph hypothesis, they should be CdC r , the rest CaC a ; 9 gave only dilutes and red-eyes (18 dilutes, 24 red-eyes); 9 others gave only dilutes and albinos (20 dilutes, 16 albinos). There were 3 which had had only 8 dilute young when tabulated. The fact that none of the 9 which had red-eye young also had albinos among 42 young gives a third body of evidence pointing toward the allelomorph hypothesis. These results make it reasonably certain that the allelomorph hypo- thesis is correct. The only other possibility would involve coupling so close as to simulate multiple allelomorphs. The hypothesis of allelo- morphs has been reached by a method of elimination. It remains to show that all of the data are in harmony with it. In the next section, definite conclusions are reached as to the inheritance of variations in intensity and dilution which make it possible to distinguish intense animals from dilute in all but very exceptional cases. 84 INHERITANCE IN GUINEA-PIGS. The following table gives a summary of the data bearing on the inheritance of the albino series of allelomorphs based on these conclu- sions. It will be noticed that animals of every possible formula have been tested by crosses with albinos, the lowest recessives. No attempt has been made to make all other possible crosses, and several (especially TABLE 36. Summary of albino series crosses (crosses 16 to 44)- Parents. Formulae. Int. Dil. R.E. W. From crosses Intense X albino . . . CC X C a Ca 40 17a 6. 31 36 186, 34, 41 9 4 21 64 71 17c, 17d, 18c, 22 Dilute X albino CdCdX C a C a ... 26 16a, 38a, 44 18 24 19,27 Red-eye X albino CdCa CaCa- 98 30 79 166, 16c, 19, 27, 33, 386, 44 25 6 3 23, 25 Albino X albino x Long established. Intense X red-eve cc x( r r l.. 9 20 ICrCaJ ccd /S r S r V- 28 31 18a, 20 Dilute X red-eye (.C-r^aJ cc - & CdCdx( r r V. 25 15 20 7 18c 20 26 43 |_UrO a J 1 1 20 13 5 g 26 43 Red-eye X red-eye 17 6 24 Intense X dilute CC X CdCd. . 14 28 CC CdC a . . CCd CdCd . . 12 in 7 35 29, 40o CCd CdC a . . 39! 40 32, 36, 40a CC a CdCd - . 191 10 28, 40o Dilute X dilute . ....... C/C_/a ^d^-'a- CdCdX CdC a ... 28 22 15 15 36, 406 30 37 42 Intense X intense CdCa CdCa- CCd X CCd . . 57 82 19 24 37,42 31 39 CCd CC a . - - 75 35 39 CC CC x See rough and Lima crosses. ones involving red-eye) have not yet been made by the writer. The last column refers to the crosses tabulated at the end of the paper. The ratios expected are obvious from the nature of the matings, except that 39 to 44 were not random crosses of their kind. The appearance of recessive young was used as a criterion of the nature of the parents in these cases. This causes an excess of recessives to be expected. INHERITANCE OF MINOR VARIATIONS IN INTENSITY. 85 INHERITANCE OF MINOR VARIATIONS IN INTENSITY. METHODS AND ACCURACY OF GRADING. The method of grading has been described on page 60. Every guinea-pig which showed dilute black or yellow in the fur was compared with standard samples of hair within a week of birth. These samples were black , sepia 3 , sepiag, and sepiag, in the black series, and red , yellow 3 , and creams in the yellow series. Intermediate grades were given by estimate. Grades were taken later in life in many cases hi order to determine the relation of age to intensity of pigmentation. In interpreting the results, it is important to know the accuracy with which the grading could be done and the difficulties met. In some cases the back and belly are f airly uniform in intensity, but usually the belly is considerably the lighter. Tufts of hair for grading have always been taken as near the middle of the back as possible. In some cases the hair is of fairly uniform intensity from base to tip. In most cases, however, the base is very much lighter than the tip. The color at the tip has been used in grading, although extreme varia- tions in the intensity at the base have also been noted. The color at the tip has most to do with the general appearance of the animal. The attempt has been made to get both a yellow and a sepia grade for every animal, so that the correlation between the intensities in these series could be determined. This is easy in the sepia and yellow-spotted animals, but in agoutis (where the yellow band of the agouti pattern displaces the sepia near the tip of the hair) determination of the inten- sity of sepia has not been so satisfactory. Several independent determi- nations have been taken in many of these cases. In most cases the same grade was assigned the second time and rarely did the second grade differ from the first by more than one point. VARIATIONS IN INTENSE GUINEA-PIGS AND ALBINOS. Before discussing the inheritance of variations among dilutes, it will be well to note briefly the range of variation among guinea-pigs which have the intensity factors (CP). In the BW race the blacks are a very intense black. The base of the hair is only slightly lighter than the tip. In other races, especially the 4-toe stock, the tip of the hair is a dull slaty black and the base a very dull color, often with less pigment than many typical dilutes. The animals have a dull streaky black appearance very different usually from the uniform dark sepia of the darker dilutes. This dull color is not associated with heterozygous albinism. Male M330 was undoubtedly homozygous (CC), having had 9 intense young by albino females and no others; yet he was one of the dullest blacks in stock. On the other hand, nearly all of the intense blacks of the BW race are heterozygous for albinism. 86 INHERITANCE IN GUINEA-PIGS. This dull black can not be due to an allelomorph of albinism between intensity (C) and dilution (C d ), since it is a condition which can be transmitted by albinos. Indeed, the albinos themselves of the BW and 4-toe stocks differ conspicuously in appearance. The BW albinos have jet black ears and feet, dark smudges on- the nose, and usually some sootiness on the back. The 4-toe and most other albinos (at the Bussey Institution) have very much less black on ears, nose, and feet, and the rest of the fur is pure white. There are parallel variations hi the intensity of red in these stocks. The occasional red spots in the BW race are of a very intense red (standard redo). In the 4-toe and other dull stocks the red is consider- ably less intense, especially at the base of the hair. The most dilute grade found hi tame guinea-pigs known to have factor C is yellow 2 (D12 cross 35-1). The wild Cavia cutleri is quite light in color. The black of the fur is a dull slaty color, more like the dull black of the 4-toes than any other color in tame guinea-pigs. The yellow on the back is about yellow 3 , on the belly creams. In spite of the resemblance to tame yellow agoutis, Cavia cutleri has the intensity and not the dilution factor. When crossed with animals of the BW race, whether blacks or albinos, the young are intermediate in intensity and would be called intense (Parti). Crossed with black animals of the 4-toe race, the young are but little more intense than Cavia cutleri. (See plate 3.) Summing up : All variations maybe found among intense guinea-pigs, from uniform black to a dull slaty black 2 and from red to yellow 2 . In the dull grades the hair is especially dull at the base. These variations are hereditary, but have not been analyzed. The hereditary factors for these variations in intense guinea-pigs are responsible for visible differences among albinos. It is to be expected, as indeed is the case, that variations will be found among dilutes, for which these same unanalyzed hereditary differences of different stocks are responsible. Finally, the residual heredity of all tame guinea-pigs has more intensi- fying effect than that of Cavia cutleri, the wild species. MULTIPLE ALLELOMORPHS. The presence of at least four allelomorphs in the albino series suggests the hypothesis that other allelomorphs in the series may be responsible for the intermediate grades hi intensity. It is a tempting hypothesis to suppose that the continuous series of variations is correlated with a continuous series of allelomorphs, such that each grade of intensity is dominant over all lower grades. If this were the case a stock of dilutes, hi which all derive their dilution from a single gamete of one animal, should be fairly constant hi their degree of dilution. Again, the cross of dilute by dilute should never give young more intense than the darker parent. INHERITANCE OF MINOR VARIATIONS IN INTENSITY. 87 However, both of these tests fail. No single gamete stock of dilutes has been found which will not give the entire range of variation when tested. Thus, male D30 redo, an intense which carried dilution as a recessive (CCd), was crossed with red-eyes. His dilute young must all owe their dilution to the same single gamete. They ranged from D340 blacki to D 152 sepia 5 . Yellows which owe their dilution to this same single gamete (derived from male 00 creame CaC a , the father of D30) range from D391 yellow 2 to 00 creamy . Dilution from a single gamete of A674 sepiae (CdC a ) has given rise to D652 black! and M306 sepia?, D409 yellow 3 , and M199 cream 7 . This last case involves no admixture of South American blood. Inspection of the tables will yield many similar cases. Evidently dilution from a single gamete may appear hi dilutes of any grade of intensity. The extreme variations may occur within a single litter (offspring of D30). Again, many examples can be given in which the offspring are much darker than either parent. D652 blacki was the offspring of D215 sepia 3 and D106 sepia 4 . In cross 37 there are 6 cases in which creame X cream^ has produced yellow 3 , with other less extreme cases of this kind. These results do not demonstrate that no more than four allelomorphs in the albino series are present in our stock. They do show that there are other causes producing varia- tion of much more importance than any other allelomorphs which may be present. THE RELATIONS OF IMPERFECT DOMINANCE, STOCK, AND AGE TO GRADES OF INTENSITY. In tables 37 and 38, and diagrammatically in figure 5 and figure 6, all records of grades of dilution at birth are analyzed with respect to genetic constitution and stock. All of those whose genetic constitu- tion was known with complete or nearly complete certainty, either from parentage or from offspring, are put after the proper formula, CdCd, CaC r , etc. All from litters containing two classes are listed separately with the numerical expectation of the classes as (20 CdCd : 32 CdC a ), etc. Those hi the litters whose formulae were later determined by a test mating are given below in parentheses. These tested individuals are included both among those of certain constitution and hi the litters containing two classes. No very close analysis of the influence of stock was possible from the data obtained. However, the following stocks were recognized: Dil., Dilute selection stock. Misc., Miscellaneous stocks with but little BW blood and no lea or Arequipa blood. These contained much dilute selection and 4-toe blood and some C. rufescens ancestry. $BW, FI from the cross of miscellaneous with BW stock, f BW, Back-cross of $BW with BW stock. S. Am., All animals with lea or Arequipa blood, in most cases about J South American, J BW, and i miscellaneous, but including pure lea, IcaXBW, etc. 88 INHERITANCE IN GUINEA-PIGS. In each array of animals of known constitution and stock the number of animals involved, the mean grade of dilution, and the standard deviation of the frequency polygon are given. It will be noticed that the standard deviations decrease as the analysis is made closer. For Red Red, White FIG. 5. Variations in intensity of yellow. Formula and stock printed near mode of each distribution. example, the standard deviation for all dilute blacks is 1.53, for dilute blacks of formula CdC a is 1.13, and for those of formula CdC a and of South American stock is 1.02. The corresponding numbers for dilute yellows are 1.10, 0.76, and 0.59, respectively. INHERITANCE OF MINOR VARIATIONS IN INTENSITY. 89 In tables 39 and 40 are given the mean grades at birth and when more than 4 months old for all guinea-pigs which were graded these two times. These data are arranged by constitution and stock. In most cases the mean grade at birth of the sample graded twice agrees well with the mean grade at birth of the whole array of the same constitu- tion and stock. BlacK Se Pa -Variations in intensity of dilute blacks. Formula and stock printed near mode of each distribution. VARIATIONS OF YELLOW. I owe the suggestion that heterozygous albinism may be correlated with extreme dilution of yellow to Professor Castle, who found that attempts to select for a cream stock of maximum dilution led to stocks which invariably gave numerous albinos. The tables and figures con- firm this suggestion in a very striking way. Animals known to be homozygous dilute (CdCa) vary between yellow 2 and yellow 4 with the mean at yellow 2 .9. Those known to transmit albinism vary between 90 INHERITANCE IN GUINEA-PIGS. TABLE 37. Effects of stock and imperfect dominance on intensity of yellow. Constitution. Stock. Redo Redi Y 2 Y, Y 4 Cr 6 Cr 6 Cr, No. Mean. <7 Y Cr CdCd Dil 1 3 10 1 2 1 9 2 3 1 3 27 19 (3) (2) 1 7 (1) 2 11 12 13 37 55 8 12 57 3.50 2.64 3.00 4.31 5.51 5.54 4.88 4.75 4.58 0.50 .77 .41 .46 .60 .71 .33 .43 .59 Do Misc 6 1 6 Do S.Am CdCr ...Do 4 14 23 7 9 27 20 (3) (4) 13 19 CdCa Dil 21 25 4 Do Misc 1 33 Do JEW Do IBW Do S.Am . . 3 1 1 4 (25 CdCr: 15 CdC a . iCdCr by test . ..Do ...Do ( CdCa by test . ..Do 12CdCd:23CdC a . (29 CdCd: 43 CdCa. i CdCd by test Dil 9 20 (2) 12 18 Misc 3 (1) 5 4 CD 5 . ..Do [CdCa by test . ..Do (4) 3 (3) (2) (2) 2.5 CdCd: 3.5 CdCa. CdCd S.Am.BW. Total 3 7 14 4 9 36 70 25 13 169 335 2.88 4.31 5.12 4.72 .65 .46 .76 1.10 6 CdCr . ..Do 4 80 128 CdCa . ..Do 49 77 4 7 1 10 34 41 Dil . ..Do 9 44 TABLE 38. Effects of stock and imperfect dominance on intensity of black. Constitution. Stock. Bo Si 82 S 3 S 4 s S 6 87 S 8 No. Mean. <7 CdCd Misc 1 6 3 1 11 2 45 99 Do S.Am. . 6 4 10 1.40 .50 CdCr . . Do . . . 6 7 ?, 1 1 17 2.06 .94 CrCr ..Do.. . 9, a 4 1.00 1 00 CdCa Misc. . . ? 4 q ?,3 3 41 5.51 1 05 Do iBW . 7 11 17 1 36 4.33 82 Do xBW i q 2 3 15 3.47 88 Do S.Am. . . . Do . . 2 18 16 37 9 4 19 17 7 17 2 4 3 85 81 4.20 4 73 1.02 1 32 (20 CdCd: 32 CdCa Misc. . . q 9 6 8 13 7 jCdCd by test . . Do . . . C? 1 ) [ CdCa by test . . Do . . . CD CD C3) C?,) 4 CdCd: 8 CdCa iBW .. 1 i q 1 1.5 CdCd: 1.5 CdCa S.Am. . 1 i 1 CdCa bv test . . Do . . . CD 29 CdCr: 20 CdCa ..Do... 10 6 9 1?, 4 6 ?, CdCr by test ..Do... Cfil CD C?r) CD CD CdC a by test . .Do. C 9 ) (2) C?t C 9 '* 6 CrCr: 11 CrCa . . Do . . . ?, 5 ?, 3 3 ?, CrCrby test . .Do. . . C?) CD CrC a by test . .Do. . . CD CdCd Total. . 7 10 3 1 ?1 1.95 95 CdCr . .Do.. . 6 7 a 1 1 17 2.06 .94 CrCr . .Do. . . ?, 4 1.00 1.00 CdCa . .Do.. . 3 36 f>4 48 31 5 177 4.47 1.13 ..Do. . 16 94 17 17 4 3 81 4.73 1.32 CdrCdr . . Do . . 9 13 19 5 1 ? 4? 1.90 1 06 CdrCa . . Do . . Ft 5? 78 65 48 Q 3 9 58 4.55 1 20 Dilute . . Do . . >n 34 66 70 58 45 10 303 3.95 1 53 Red-eye Dil + RE ..Do... ..Do... 2 2 20 6 40 18 84 26 96 20 78 19 64 4 14 3 3 98 401 4.45 4.07 1.54 1.53 INHERITANCE OF MINOR VARIATIONS IN INTENSITY. 91 yellow 4 and cream 7 , mean at creams.! very distinctly paler. Litters which should give both have given the entire range with two modes, at yellow 3 and cream 5 , respectively. It is especially to be noted that among 13 of these, which were given grades before their constitution was known, 4 ranging from yellow 2 to yellow 4 proved to be homozygotes, while 9 ranging from cream 5 to cream 7 proved to be heterozygotes. Dilutes known to transmit red-eye (CdC r ) have been either yellow 4 or cream.5, mean at yellow 4 . 3 . These should be compared with those of TABLE 39. Effect of age on intensity of yellow. Constitu- tion. Stock. Mean. No. in sample. Mean at birth. Mean adult. Dark- ening. CdCd Misc-Dil 2.8 9 3.1 2.9 0.2 CdCa Misc .... 5.5 17 5.1 5.0 .1 Do Do Do CdCr Dil J-f BW . S.Am . . . . . Do 5.5 4.8 4.6 4.3 9 11 9 5 5.2 5.0 4.4 4.8 6.0 4.7 4.7 4.2 - .8 .3 - .3 .6 TABLE 40. Effect of age on intensity of black. Constitu- tion. Stock. Mean. No. in sample. Mean at birth. Mean adult. Dark- ening. CdCd CdC a Misc .... . .Do 2.5 5.5 8 14 3.0 5.6 2.4 4.6 0.6 1.0 Do Do Do BW.... fBW.... S.Am . . . . . Do 4.3 3.5 4.2 4.7 20 6 15 8 4.3 3.3 4.8 4.9 3.2 2.5 3.3 2.0 1.1 .8 1.5 2.9 CdCr . .Do 2.1 16 2.2 1.1 1.1 CrCr . .Do 1.0 4 1.0 1.0 the same stock (S. Am.) which transmit albinism. The difference, yellow 4 . 3 compared with yellow 4 . 6 , is too small to be relied on. Litters which should give both CdC r and C 69 C, D Tri Sm pure cut 1 1 9 Total 5 2 15 Expectation Ss + SS ( 1 ] 1 11 Cross. SS XSS A B C D E 55 E Tri E Tri 4 Expectation SS ( 1 ( 1 4 The interpretation given is no doubt open to objections. In some cases the ratios seem rather aberrant. This is in part due to the small numbers, but also to the overlapping of class ranges. In most cases rough B must be considered as full-rough genetically (Rss), but in some cases it is probably partial-rough (RSs) . Rough E usually seems to be RSS, but in some cases must be heterozygous (RSs). It has not been demonstrated that factor S of the wild species is identical with the similar factor of the tricolor stock. If not identical, however, the latter stock differs from the wild by two mutations which neutralize each other, while if identical we can consider that the original tricolor stock had simply persisted in the primitive condition, never having had the rough intensifying mutation, s, of the fancier's roughs. MINOR VARIATIONS. Probably part of the minor variations in roughness are due to chance irregularities in development which are not hereditary. This is indi- cated by the slight asymmetry not uncommonly present. This asym- metry seldom amounts to more than the absence of a member of one pair of rosettes. No Mendelian analysis has yet been attempted for minor variations, but certain hereditary differences between different stocks are quite ROUGH FUR. 117 clear. The Lima stock shows a distinctly lower level of development of roughness than is found in the 4-toe stock or even among the full- roughs of tricolor stock. A large part of the variation and overlapping in the remaining experiments in which various stocks have been mixed is made intelligible by assuming that the residual heredity is unfavor- able for roughness in the wild species and especially favorable in the 4-toe stock. If we let S+ stand for favorable and S for unfavorable residual heredity, the wild species and presumably the primitive guinea-pigs are rrSSS , while the good fancier's roughs, RRssS-j- differ by at least three independent sets of factors, all favorable for roughness. ROUGHNESS OF SERIES II. It has been mentioned that irregularities in hair direction have been found in certain stocks which can not be classified by the grades which have been defined. The BW race is a highly inbred race. No indi- viduals of the pure stock have ever been observed to have roughness on the face, back, or toes, but many of them show irregular partings and crests along the chest and belly. It will be remembered that hi series I ventral roughness appears only in high-grade roughs grades A or B. Thus the characteristic roughness of the BW stock is nearly the least characteristic feature of series I. The only distinction which has been made in these BW roughs is between strong-rough with two or more ridges and poor-rough with only one ridge or a mere trace of roughness. Table 61 shows the principal results. TABLE 61. Smooth. Poor rough. Strong rough. Smooth X smooth Poor X poor 11 14 6 1 6 1 Strong X strong ... . 5 5 16 It is clear that this roughness is due neither to a simple dominant nor to a simple recessive. Aside from this, the results are exceedingly difficult to interpret, since poor X poor gives more smooth than does smooth by smooth. Probably the results will become more harmonious when more data are obtained. It seems safe to conclude at present that this roughness is wholly independent of ordinary roughness in its causation. Irregularity in hair direction on the back, not resembling anything hi series I and not correlated with roughness of the hind toes, has been observed in a few individuals of Arequipa and lea stock. It does not seem to be like the BW roughness, but resembles the latter in the irregularity of its inheritance. 118 INHERITANCE IN GUINEA-PIGS. SUMMARY. The principal results which have been reached may be summarized as follows: 1. A classification of guinea-pig fur, skin, and eye colors is given with definitions of fur colors in terms of Ridgway's charts (1912). 2. Rodent color factors are conveniently classified as follows : a. Factors which affect the distribution and intensity of color largely irrespective of the kind of color. b. Factors which govern the differentiation between yellow and dark colors in colored areas of the fur. c. Factors which determine the kind of dark color in the areas with dark pigmenta- tion in fur and eyes, without influence on yellow areas. Definitions of all known guinea-pig color factors are given on this basis and a table of the color varieties arising from combinations of these factors is given. 3. Genetic and biochemical evidence on the physiology of pigment formation suggests the hypothesis that the three groups of factors determine respectively the distribution and rate of production by the nucleus of the following substances : o. A peroxidase which, acting alone, oxidizes chromogen in the cytoplasm to a yellow pigment but is so unstable that it must be produced at a relatively high rate to give any pigment at all. b. A supplementary substance which, united with the first, makes it a dark-pig- ment-producing enzyme and of such stability that color develops at a much lower level of production of peroxidase than when the supplement is absent. Above the level at which both produce effects, the dark and yellow-producing enzymes compete in the oxidation of chromogen. e. Additions to the second substance which cause variations in dark color but not in yellow or in the competition between dark color and yellow. 4. There is a continuous series of variations in intensity of pigmen- tation hi the yellow, brown, and black series and in eye color. The ordinary dilute guinea-pigs are found to be imperfect albinos in the sense that dilution is due primarily to a member of the series of allelo- morphs intensity, dark-eyed dilution, red-eyed dilution, and albinism, with dominance in the order of increasing intensity. 5. A further step in the analysis of the continuous series of variations of intensity is taken hi the demonstration that dilution is imperfectly dominant over red-eye and albinism as regards the yellow series of colors, and that dilution and red-eye are imperfectly dominant over albinism, as regards the black series. Smaller effects are due to the residual heredity of different stocks and to age. 6. Evidence is presented which confirms the hypothesis of Detlefsen (1914) that the light-bellied agouti pattern of tame guinea-pigs, the ticked-bellied agouti of hybrids between the tame guinea-pig and Cavia rufescens, and non-agouti (as seen in self blacks or browns) form a series of triple allelomorphs in which light-belly is the highest dominant and non-agouti the lowest recessive. Evidence is presented which GENERAL CONCLUSION. 119 indicates that Cavia cutleri possesses the same agouti factor as tame agouti guinea-pigs. Light agouti of Cavia cutleri and dark agouti of Cavia rufescens are thus variations in a character in two wild species which differ in heredity by a clear-cut Mendelian factor. 1 7. There is a continuous series of variations between smooth fur and very rough or resetted fur in guinea-pigs. The primary effects in this series are due to two independent pairs of allelomorphs. One factor, discovered by Castle (1905), is essential to any roughness of the common type, and is completely dominant over its allelomorph found in wild cavies and smooth guinea-pigs; the other, an incomplete recessive to its allelomorph in the wild cavies and some tame guinea-pigs, is necessary for the higher grades of roughness. Second-order effects seem to be due to the residual heredity of different stocks, and probably to non-hereditary irregularities hi development. There is a roughness of a different type from the usual which is inherited independently. GENERAL CONCLUSION. Most of the successful earlier attempts at Mendelian analysis of heredity naturally dealt with variations which were obviously dis- continuous. But in nature such variations are much less common than apparently continuous series of variations. It was thus a common reproach against the Mendelian analysis that it dealt only with excep- tional conditions. The work of Nilsson-Ehle, East, and others has shown how quantitative variation may be brought under a Mendelian explanation. MacDowell (1914) presents data on size inheritance from this standpoint and discusses the literature up to that time. Recently two very interesting papers have been published (Dexter, 1914, Hoge, 1915) which analyze the heredity of certain very variable char- acters in Drosophila by means of linkage relations. Several of the studies in this paper deal with inheritance in continu- ous series of variations. The only general statement which can be made about the results is that there is no general rule for such cases. Intermediates between varieties which mendelize regularly have been found to follow very definite modes of inheritance, which, however, are very different in different cases and could not possibly be predicted a priori. On the other hand, each mode of inheritance is exactly paralleled by cases among the most diverse groups of animals and plants. It may be interesting to summarize the modes of inheritance of inter- mediates which have been found. An intermediate condition is sometimes found to be due to an inter- mediate variation of the essential hereditary factor involved, i. e., to an allelomorph. Thus yellows are intermediate between red and albino 1 It should be pointed out, however, that the original stock of Cavia rufescens used in these experi- ments included individuals of the light-agouti character as well as those classed as dark agouti. It seems quite likely that dark agouti arose as a recessive mutation in C. rufescens. W. E. C. 120 INHERITANCE IN GUINEA-PIGS. guinea-pigs in appearance, and we find an allelomorph intermediate in dominance between the intensity and albino factor to be responsible for their condition. Sepias are similarly intermediate between blacks and albinos and are due to the same allelomorph of intensity and albin- ism. The series, light agouti of Cavia cutleri, dark agouti of C. rufes- cens, and black, furnishes another example due to triple allelomorphs. In other cases, the intermediate type is an unfixable one, due to imperfect dominance. Thus cream is the heterozygote between yellow and albino. A "razor back" rough (rough C or D) is the heterozygote between a type smooth except for the hind toes (rough E) and a full- rough (rough A). A series of deviations from the original type may depend on the presence of a certain factor necessary for any deviation whose effect is modified to different extents by independently inherited factors. Rough A contains the same rough factor (R) as does rough E, but differs in possessing an independent factor variation (s) favorable for rough- ness. Most of the variation which we have ascribed to residual heredity probably comes under this head. Deviations from type, which apparently form a natural series, may be due to wholly independent factors whose effects are merely super- ficially similar. A pink-eyed pale sepia superficially seems as good an intermediate between an intense black and an albino as does a black- eyed sepia, yet the former is due to a variation which is wholly inde- pendent of albinism; the latter is due to an allelomorph of albinism. White-spotted animals are sometimes called partial albinos and con- sidered as natural intermediates between the self-colored type and albinos, but genetically they are wholly distinct. Black, agouti, and self yellow form a series which is due to three allelomorphs in mice, but in guinea-pigs two wholly independent sets of factors are involved. Finally, we must recognize series of variations in which no Mendelian factors have yet been isolated. The series of white-spotted and yellow- spotted types and the series of polydactylous types are examples hi guinea-pigs. Further, in all series of variations, to whatever extent analysis has been carried, there always remains some unanalyzed varia- tion. In many cases such variations are known to be hereditary and can be assigned to the residual heredity of particular stocks. Such unanalyzed variations, however, are probably in general complicated by variation which is not hereditary, due apparently to irregularities hi development. If we can measure the importance of such non- hereditary variation by the extent of irregular asymmetry met with, it is very important in white and yellow spotting, in the variations in the development of extra toes on the hind feet, and is noticeable in varia- tions in roughness. In the continuous series of variations several of these, phenomena have generally been found together. In the series from smooth to full- TABLES. 121 rough we find a primary unit difference, a modifying factor, imperfect dominance in the effects of the latter, effects of residual heredity, and probably some non-heritable variation. In the series from red through yellow and cream to white we find multiple allelomorphs, imperfect dominance, and small effects due to residual heredity. In the series black through sepia to white, we find independent factors, multiple allelomorphs which show imperfect dominance, and rather prominent effects due to residual heredity and to age. This last series is interesting as at least a close parallel in appearance to the series of variations in human hair black, brown, tow-color, to white. Thus hi each case a complex of the most varied causes underlies an apparently simple continuous series of variations. EXPERIMENTAL DATA. EXPLANATION OF TABLES 62 TO 137. Crosses 1 to 15 include all matings recorded by the writer which involve the inheritance of agouti and in which at least one of the parents had Cavia rufescens ancestry. A large part of the remaining crosses are non-agouti by non-agouti, producing only non-agouti young. All the young in which the agouti factor should produce a recognizable effect, if present, are classified under the heads Lb, Tb, and Non, which mean light-bellied agouti, ticked-bellied agouti, and non-agouti, respectively. Most of these are the typical (black-red) light-bellied or ticked-bellied agouti or black. Those which are not typical, e. g., brown-red agouti light-belly, red-eyed sepia, etc., are described further under the column " Remarks." Those young in which the agouti factor can produce no visible effect, even though present (albinos, reds, yellows, and creams), are described under the column " Unclassified." Thus the exact color of every one of the young from each mating can be found from the tables, with the exception that white and red spotting are not noted. The matings in each cross are numbered in the first column. The number, description, and descent of the mother and father are given in the second and third columns, respectively. As in the case of the young, black-red agouti light-belly or ticked-belly or black, depending on the heading of the column, are understood where no description is given. The descent is indicated in most cases by a reference to the mating from which the animal was derived. Thus 36-4 means mating 4 of cross 36. In other cases the stock is indicated as BB or BW. The symbol ArF 2 means F 2 from crosses of Arequipa c? 1002 with guinea- pigs. In some cases merely the amount of Cavia rufescens blood is given. Thus M49, in the first cross given, was an ordinary ticked- bellied agouti from mating 9 of cross la. Referring to this mating, we see that his parents were female 84, a black of BB stock, and male A1121, a ticked-bellied agouti with ^ Cavia rufescens blood. 122 INHERITANCE IN GUINEA-PIGS. Crosses 16 to 44 include all matings recorded by the writer in which there was dilution or red-eye in either parents or offspring, except for a few cases among the Cavia cutleri hybrids and cases of intense by dilute with only intense young. Some other crosses are included for special reasons bearing on the inheritance in the albino series. There is some repetition from matings outside of 16 to 44, but most of those outside are intense by intense, with only intense and, in some cases, albino young. As in the agouti crosses, all matings are numbered in column 1. The number, description, and descent of the mother and father are given in columns 2 and 3, respectively. All the offspring are classified under the heads Int, Dil, RE, or W, which stand for intense, dilute, red-eye, and white (albino), respectively. A further description of all except the albinos is given under the column " Remarks." The attempt has been made to give the grade of dilution at birth for every dilute or red-eyed animal where known. Crosses 45 to 57 give the data on the inheritance of rough fur in the 4-toe and tricolor stocks. As before, the matings are numbered. The young are classified under the heads A, B, C, D, E, and Sm, which refer to the grades of roughness defined in the paper and to smooth. The parents and offspring were black (usually with red and white blotches) except for a few cases which are all noted. Such a symbol as red-B means a red of grade rough B. Crosses 58 to 62 give the results in the pure Lima stock and 63 the results in the cross of Lima with other stocks. Where no color is given black is always to be understood. Crosses 64 to 66 give the matings involving rough fur among Cavia rufescens hybrids which were recorded by the writer. Cross 67 gives crosses of pure lea with rough A stock. Crosses 68 to 78 give all the data in matings involving Cavia cutleri ancestry made by the writer. The following symbols are used : AgLborAg = Black-red agouti, light-belly. AgTb = Black-red agouti,ticked-belly. B = Black. BrAgLb = Brown-red agouti, light-belly. BrAgTb = Brown-red agouti, ticked belly. Br = Brown. R = Red (black-eye). R(Br) = Red (brown-eye). SYAgLb = Sepia-yellow agouti, light- bdly. = Sepia-yellow agouti, ticked- belly. = Sepia. SYAgTb Sep, S SAg(R) Sep(R) W Red(p) Sep (p) In such expressions as S 3 Y 3 Ag the numerals stand for the grades defined in the text. In crosses 1-15, Lb and Tb are used at the heads of the columns to include any light-bellied or ticked-bellied agouti. Non means non-agouti. A, B, C, D, E, and Sm are used for grades of roughness and for smooth. BrYAgLb = Brown-yellow agouti, light- belly. BrYAgTb = Brown-yellow agouti, ticked- belly. LBr = Light brown. Y = Yellow (black-eye). Y(Br) = Yellow (brown-eye). Cr = Cream, used in compounds likeY. = Sepia-white agouti (red-eye). = Sepia (red-eye). = White or albino. = Red (pink-eye). = Sepia (pink-eye). TABLES. 123 TABLE 62. Cross 1. Matings of non-agouti (aa) with ticked-bellied agouti (A 'a). Each of the latter known to be heterozygous because of a non-agouti parent. Expectation: A'a X aa = A'a + aa (1 AgTb : 1 Non-Ag). la. Mother non-agouti, without rufescens ancestry. No. 9 Non-Ag. rfAgTb. Lb. Tb *Jon Remarks. Unclas- sified. 1 2 3 4 6 6 7 8 9 10 11 12 13 14 15 16 17 18 399 BW... 499 4-toe. . M49 la Do -q 9 5 3 2 4 1 6 9 1 9 8 9 2 3 2 4 4 2 5 6 W 1 W 5 W 1 W 2 W 399 BW. . . 399 4-toe. B5 ld-16. Do ...---- 299 BW. . . 65 BB . . . 299 BW. . . 399 BW. . . 84 BB . . . C22 Misc.. C35 Misc.. 399 4-toe. . 3W BW... 3 W Misc . . D44Sep 16o-3. AA244 Sep 2-12 B27 Id B30 la B69 la B191 la A1121 & A1474 ft Do -14. -1 -1 2 W 4 W -5 ? B171 la B117SCrAgTb Id Do -4 8 3 2 1 6 2 3 1 W 5 W 1 W 2 W -11. 3 SCrAgTb, SCrAgTb.Bi 3 Sep. SCrAgTb 2 Sep CrAgTb, Do Do 1 SCrAgTb BW 43 W BW. . . Sep (R) S.Am . Total D113BCrAgTb 3a-7 . AA433a 3b-A . 1 4 1 2 SCrAgTb, S 3 AgTb, SAg Sep(R). p Tb(R), 2 1 W 62 62 Ib. Mother non-agouti, with rufescens ancestry. No. 9 Non-Ag. cfAgTb. Lb Tb Non Remarks. Unclassified. 1 2 3 4 5 6 7 8 9 10 11 A443 A469 | 1 1 LI tr A1390 A A1227W ^ /A1413 & \A1291W A A1309 W tf A1407 5 ! , A1413 ^ M115W ^ fM114 16-7... A1050 g'j A781 Jz.. 3 SC rAgTb ^A1449 T& . . 1 1 3 3 1 2 1 2 1 W R, R(Br), Cr(Br) A1513 & Br Br AgTb A1449 s>j Do M189 lc-3 1 Do.. 4 1 4 1 SC Br rAgTb, Sep . . \M90 Br M90Br jfe / Do CrAgTb, Sep Cr W A1330 its---- Total A1331 f iff.. 17 13 124 INHERITANCE IN GUINEA-PIGS. TABLE 62 Continued. Ic. Male genetically, but not visibly ticked-bellied agouti. No. 9 Non-Ag. tfAgTb. Lb Tb Noi i Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 131 W G.p.. 13a G.p.. 20 G.p.. 58 Sep Dil . . 17, 30Cr(B) Dil . A412 R(] Do Br) Ar. 2 3 Se p Do 3 3 1 1 SC 3 Se SC SC *Ag1 5CrA p. . . rb B42 W la-3 . . Do ? Tb 2 W W 55 Cr(Br) Dil . Do . 1 *AgT *Ag1 "b M292 &... M326 jV D18W Do lc-5 . . 2 1 1 3 1 1 :b, 2 Sep. . . 4 W M353 fa Do Se p Total 11 12 Id. Male non-agouti, without rufescens ancestry. No. 9 AgTb. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 ^606 AgLb J . 166 4-to 2966 BB 3013 BB Do e ? IA450 AgLb i . M 1 A340 & A341 & A357 ^ A1146 & A1058 sS A1171 & A1058, A1171 sfr Do 4 6 1 7 1 3 5 2 6 1 2 ... Do . . Do Do Do 2996 BB 1357 BW . . . .Do . 3 5 2 1 2 5 1 2 W A1117 ^ Do SCrAgTb.... 2996 BB 3013 BB 1357 BW Do Do 2 2 3 1 1 9 1 4 1 1 1 2 11 A1450 s 1 , A1117, A1450 & A1582 ^ A1583 ^j A1677 & A1678 S V B8 Id- B23 Id- B26 Id- M113 Ib- M442 BrCrAgTb 11 Total W 2996 BB Do Do . . Do .... -7 . . Do 3 5 2 2 2 7 1 1 2 -12 Do -14 Do -7 C20 Mis 86 WBW ( j-10 2 Sep l l 61 63 TABLES. 125 TABLE 62 Continued. le. Male non-agouti (genetically) with rufescens ancestry. No. 9AgTb. c? Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 A1146 fa A504 W fa 2 2 Sep B132 ld-3 M293 Y 42-14 2 SCrAgTb B95 ld-4 Do 1 B24 ld-4 Do 1 B52 la-3 M201 W 42-13 .... 2 B33 ld-18 Do 2 W W W B110 lo-l Do Bill lo-l .Do ... 1 1 2 B128 lo-l Do B23 ld-12 Do 1 Total 10 5 , but agouti known to be derived from C. rufescens. SUMMAKY OF CROSS 1. No. M A'a Lb Tb Non la 9 9 Non-Ag (g.p.) .... cf d"AgTb 62 62 16 9 9 Non-Ag (hybrid) . . cfcfAgTb 17 13 le 9 9 Non-Ag d" cTA'a (R or W) . . . 11 12 Id d" cT Non-Ag (g.p.) .... 9 9 AgTb 1 61 63 le d" c?aa(Y or W) hybrid 9 9 AgTb 10 5 Total ~1~ 161 155 TABLE 63. Cross 2. Matings of ticked-bellied agouti (A'a) with ticked-bellied agouti (A'a), in which both are known to be heterozygous because of the parentage in each case. Expectation: A'a X A'a = A'A' + 2A'a + aa (3AgTb : 1 Non-Ag). No. 9 AgTb. cT AgTb. Lb Tb Non Remarks. Unclas- sified. 1 B15 ld-6.. B118 ld-6.. 1 2 2 B58 ld-15. Do 9 2 3 BrAgTb, 3 B59 ld-15. Do 6 4 SYAgTb.Sep 2 BrAgTb 4 B68 la-1 . . Do 5 2 SCrAgTb 5 A529 BrAgTb fa AA15 fa.. 6 1 BrAgTb, SCr 6 A913 fa Do 5 AgTb, BrCr AgTb BrAgTb 7 A1273 SCrAgTb fa AA16 fa.. 4 2 3 SCrAgTb . . . 3W 8 Do A1121 fa 1 9 10 A780 fa---- A1306 Tiit.. A781 fa---- A 1307 6 1 BrYAgTb 3R.Y 2W 11 A1561 fa A1050 fa 3 4W 12 A1566 fa . . Do 5 1 BrAgTb, SCr 12a A1566 AA15 fa.. 2 1 AgTb, Sep SCrAgTb 13 A702 fa AA16 fa 2 2 Br . 14 A1450 fa . AA433a 36-4 1 15 A1058 fa Do W 16 A1523 fa A1449 fa . . . 2 1 BrAgTb W 17 18 AA176 41-4.. AA175 41-4.. AA177 SCrAgTb 41-i.. Do 5 1 SCrAgTb W W 19 M78 9-5 A1161 fa 2 20 MHO 16-6.. A1170 fa 1 21 D26 SCrAgTb lc-4. . D33 SCrAgTb lc-6.. W Total 66 19 126 INHERITANCE IN GUINEA-PIGS. TABLE 64. Cross 3. Matings of ticked-bellied agouti from cross 2 or 12 (A'A', 2A'a) with non-agouti (aa) made in order to test for the presence of homozygotes. Expectation: A'A' X aa = A'a (all AgTb) or A'a X aa = A'a + aa (1 AgTb : 1 Non-Ag). 3a. Heterozygous females. No. 9 AgTb. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 '4 5 *6 7 8 M19 2-16.. M203 2-19 . . AA211 2-6... AA240 12-8.. AA257 12-2.. AA285 SCrAgTb 12-7. . AA242SYAgTb 12-8.. AA240, AA242 , | 12-8 . . 7 females 393 4-toe.... 2 1 1 M116Sep 42-11. . . C20 Misc. . . . 2 A1040 & 2 2 1 1 2 1 2 1 2 1 A1040.356 &, 4-toe. 393 4-toe . I5Sep(R) 21-1 BCrAgTb, 2 Sep . A1040 & 10 11 36. Possible homozygous females. No. 9 AgTb. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 AA209 2-11.. AA212 BrAgTb 2-6... AA213 2-12.. AA217 2-7 C21 Misc. 4 C20 Misc. 2 C21 Misc 8 Do 8 AA298 2-13 . . 5 females 356 4-toe 3 25 *Not certain that both parents were heterozygous (A'a) . TABLES. 127 TABLE 64. Continued. 3c. Heterozygous males. No. 9 Non-Ag. d"AgTb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 M79 W g 1 * M77BrAgTb 2-16. 1 2 1 2 1 BrAgTb.Br... BrAgTb, Br.. . 3 W /M72 lc-2 . . . \M86 9-1 ... M42LBr 42-12.. M44 Cr 42-12.. \ Do.. / AA197 2-10... Do 3 SCrAgTb M99 42-13? . M101 42-13? Do 1 Do 2 5 1 1 3 /A1407 A JAA199 SCrAgTb 2-12 \A1413 & S6 ,K.... S15 ,1* AA223 BrYAgTb 2-9 . . . ... Do 2 2 Br A1659 , ff Do 2 2 B9 Misc... Do BrAgTb S2 Misc. . . A1665 Misc. . . AA226 2-13 . . . 1 Do 1 3 2 1 1 3 R S22, A1674 ('t 8 liis- B7 ld-7... B21 Id-Q. i Do.. / AA235 12-1 . . . M Do 2 B28 ld-14 . Do 1 W W W M168 ^j... Do M169 &... Do 1 M177 lc-2 Do 2 2 3 4 3 6 4 i 1 2 BrAgTb, LBr . SCrAgTb, Sep, LBr. Br 499 Misc. . . B28 ld-14 . . M183 ^ 2 AA241SYAgTb 12-8. . 2 AA284 12-7 2 AA299 12-8... Do SCrAgTb M261 Sep 7-7 Sep AA58 Vs Do 2 2 M261/AA58 Do 9 males 1 40 38 3d. Homozygous male. No. 9 Non-Ag. d" AgTb. Lb Tb ^on Remarks. Unclas- sified. 1 2 3 S6 t fa A, S15 ** \253 SCrAgTb 2-7 6 2 BrAgTb Dn 4 A1659 8 ^ T~)o 2 2 SCrAgTb . . . 1 tlUl l( ' 1?! *A litter with an unexpected AgLb. Paternity not wholly certain. 2 Not certain that one of parents was heterozygous (A'a). SUMMARY OF CROSS 3. AgTb, from cross A'a X A'a, tested by cross with Non-Ag. Lb. Tb. Non 3a 79 9 A'a 10 11 3b 59 9 A'A'(?) 25 3c go'd* A'a 1 40 38 3d Icf A'A' 12 128 INHERITANCE IN GUINEA-PIGS. TABLE 65. Cro 4- Ticked-bellied agoutis, known to be homozygous because of test (cross 3), or parentage (cross 4), crossed together. Expectation: A'A' X A'A' = A'A' (all AgTb). No. 9 AgTb. of AgTb. Lb Tb Non Remarks. Unclas- sified. 1 AA213 2-12 . AA253 SCrAgTb 2-7 9 2 SYAgTb, ? AA217 2-7... Do 11 SCrAgTb, BrAgTb, BrYAgTb, BrCrAgTb 3 SYAgTb, 8 AA613 4-1 . Do 3 SCrAgTb 3 SCrAgTb . . . 4 AA671 4-1 ... Do W 6 AA577 4-2... AA573 BrAgTb 4-1 .... 1 BrAgTb 6 AA606 SYAgTb 4-2. .. Do 2 SCrAgTb Total 26 TABLE 66. Cross 6. Homozygous ticked-bellied agouti (A'A') crossed with heterozygous ticked-bellied agouti (A'a) or with non-agouti (aa). Expectation: All AgTb (A'A' or A'a). No. 9 AgTb or Non-Ag. cfAgTb. Lb Tb Non Remarks. Unclas- sified. 1 9, M181 BrCrAgTb 15-15. . Do AA253 SCrAgTb 2-7... AA573 BrAgTb 4-1 1 5 SYAgTb 3 BrAgTb, W 2 W 8 M442 BrCrAgTb lb-10 . . AA573 BrAgTb 4-1 ... 1 2 BrCrAgTb BrCrAgTb... 4 M296SAgTb 14-4... Ml 16 Sep 42-11. 3 3 SYAgTb . . . 5 D194Sep(R) 26-2... AA670 SCrAgTb 4-1 ... 3 2 SCrAgTb, SAgTb(R) 2 W Total 13 TABLES. 129 TABLE 67. Cross 6. Matings of non-agouti hybrid (aa) with homozygous light-bellied agouti (AA). Expectation: AA X aa = Aa (all AgLb). 6a. Female non-agouti. No. 9 Non-Ag. cTAgLb. Lb Tb Non Remarks. Unclassified. 1 2 3 A605 J 2597 G.p ?, A642 } Do ?, A842 J Do 5 Total 9 66. Male non-agouti hybrid. No. YAgLb. cT Non-Ag. Lb Tb Non Remarks. Unclassified. 1 2 3 4 5 16a G.p . . 3520 Cr(Br) G.p . . 3o G.p . . lla G.p.. 3392 G p . . Total A674 Sep J . . . 6 Do 1040 A.. 3 7 2SCrAg,BrCrAg. Y(Br),2Cr(Br),W A504W &.. 3 A1539 ^.. 1 ?0 SCrAgLb SUMMARY OF CROSS 6. No. aa (hybrid). AA (g.p.). Lb Tb Non 6a 9 9 Non-Ag. . . d" c? Ag Lb 9 66 cT cf Non-Ag 9 9 Ag Lb. 20 Total 29 130 INHERITANCE IN GUINEA-PIGS. TABLE 68. Cross 7. Matings of non-agouti (aa) with light-bellied agouti (Aa) of rufescens ancestry, known to transmit non-agouti, because of a non-agouti parent. Expectation: Aa X aa = Aa + aa (1 AgLb : Non-Ag). la. Female AgLb. No. 9 AgLb. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 A601 -fa. . 103 4-toe... 1 1 W A614 f g . . 224 4-toe... 2 A953 d* . A718 fc ?, 3 A1310 &. 166 4-toe . . . 2 A1311 a 1 *. Do 1 1 A1324 g*z A719 W 3>j 3 W M102 66-1 Do A462W 3^ 2 2 2 Sep. M2 A J20W BW.... BW 36 W BW 3 3 SCrAgLb /M357SCrAgLb 106- \D95 SCrAgLb 106- D61 SCrAgLb' 13-5 D63 SCrAgLb 5 13-5 /D69 SCrAgLb 18-5 \M425 SCrAgLb 13-7 Total . 7 * 3 SCrAg, 2 Sep . 7 W W .....Do J....DO 2 2 2 1 2SAg(R),Sep, Sep(R) 2 SCrAg, Sep, . . W 14 18 76. Male AgLb. No. 1 9 Non-Ag. cMgLb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 67 G.p . . D43 Sep 16a-3 . M236Sep(R) ArF 2 . . D45 Sep 16a-3 . Total M123 13-2.. D94 SCrAgLb 106-8 . M331 BrCrAgLb 42-10 . D94 SCrAgLb 106-8. 2 1 1 1 SCrAg, Sep . . . W W 1 1 SCrAg, Ag.... 4 3 SUMMARY OF CROSS 7. No. Aa (hybrid). aa Lb Tb Non 7a 9 9 AgLb.... cfcf Non-Ag. . 14 18 76 3 B140 8o-2 M328B-Y 42-17.... 3 1 Do 20 W BW .... 3 2 M195 9-7 .. MllGSep 42-11... 2 2 2 Br M217 8a-3 . 356 4-toe . 1 M282 15-12 . . M116Sep 42-11.... AA83 J? . . . 2 2 2 Sep A1562 W ^i A1691 86-7 Do 6 6 7 females 17 20 106. Females AA'. No. 9 AgLb. d" Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 6 6 7 8 9 10 11 12 A389BrAgLb &.... A499BrAgLb T A1688 86-6.. A511W jg 1 2W.R R AA83 A. 1 2 1 3 2 3 5 2 2 BrAgLb.BrAgTb BrAgTb .... Do A1690 86-7.. Do 3R B139 8o-2.. Do M328B-Y 42-17.... 20 W BW 3 SYAgLb SCrAgTb B141SCrAgLb8o-2.. Do M328B-Y 42-17.... 20 W BW 356 4-toe 393 4-toe .... MllGSep 42-11.... Do 1 2 2 2 2 1 3 1 3 5 3 1 SCrAgLb, 3 SY AgTb 2 SCrAgLb, SCr AgTb 6 W M25 9-1 ... M27a 9-1 ... M82 9-7... M92 8o-4 . . SCrAgLb, BrY AgLb, SYAgTb 10 females. . . . 18 30 TABLES. 133 TABLE 71 Continued. lOc. Males Aa. No. 9 Non-Ag. ^AgLb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 75 BB . . . . 08 W 4-toe... A581 L i ? ? ~>0 2 3 4 BrAgLb 4R.Y? 09 W 4-toe... Do M7 & M133 8a-4... \ Do.. 3 2 3 SCrAgLb, Sep /M7 & \AA58 jV M183 jV / Do 1 1 M261 Sep 7a-7. . . Do 1 87 fa.... S2 ,fo . . . . M201 L > 8a-4 . . . 2 1 Do.. 4 3 R A1665 ,$ g Do 1 2 87, A1665 ,fo Do 3 2R B137 lo-l . . . /B133 ld-3... \B98 la-3. . . D148W lc-8... 5 males B155 I ] 8o-l . . . 1 2 3o.. 1 7 /' ' D240 BCrAgLbl4-5. . 1 1 SCrAg, Sep 26 33 lOd. Males AA'. No. 9 Non-Ag. ^AsLb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 31 Mis M255 lo- M253 lo- M256 la- M254, M256 la- AA279 3a-; C22, AA278 G.p B 9 9 (ab< M168 ^j. c M138 9-1 ? LO Do 4 4 1 1 1 4 5 3 1 2 1 LO Do LO Do LO Do ., 3a-3 >ve) . . Do Do Do 5 3 1 1 5 2 5 4 1 M91 8a-4... } Do.. SCrAgLb /M169 3** . . /BrYAgLb, 2 BrCr I AgTb SYAgTb 1 M171 ^j M177 Ic-S M86 9-1 M79 W j>j . / Do ' M210 15-14.. Do 3 2 1 4 1 5 3 1 1 3 2 2 2 4 2 3 BrAgLb B31 la- B53 la-; B54 la-] C29 G.p B 9 9 (alx M119 jff. L \. . .. B121 86-2... Do L. Do Do >ve) Do ... B130 8a-l... Do AA174 14-1 5 males 45 50 SUMMABT OF CBOSS 10. No. AgLb (Aa or AA') tested by cross with Non- Ag' (aa). Lb Tb Non 10a 7 females Aa 17 20 lOc 5 males Aa 26 33 Total 43 53 lOb 10 females AA' 18 30 lOd 5 males AA' 45 50 Total 63 80 134 INHERITANCE IN GUINEA-PIGS. TABLE 72. Cross 11. Light-bellied agoutis (AA') crossed together. Both parents known to carry recessive ticked-belly by test (except in the case of AA533 with one young). Expectation: AA' X AA' = AA + 2AA' + A'A' (3 AgLb : 1 AgTb). No. 9 AgLb. tfAgLb. Lb Tb Non Remarks. Unclas- sified. 1 A1690 8b-7 M91 8a-4 4 2 2 M25 9-1 Do 7 3 4 SYAgLb 3 M27a 9-1 Do 6 4 M25, M27a 9-1 Do 4 6 B139 8a-2 Do 1 6 M25, M27, B139 Do 1 3 BYAgTb 7 AA533 11-1 Do 1 8 AA588 11-2 Do 2 1 SYAgTb Total 7!5 q TABLE 73. Cross 12. Miscellaneous matings of ticked-bellied agouti with ticked-bellied agouti. No. 9 AgTb. cfAgTb. Lb Tb Non Remarks. Unclas- sified. 1 AA203 BrCrAgTb 2-5. AA16 g'j . . . 8 1 SCrAgTb, W 2 M19 2-16... Do 3 LBr W 3 AA497 SYAgTb 10d-ll AA284 12-7. 1 1 LBr R 4 AA598 3c-22 . . Do 2 BrAgTb . . . W 5 A1523 g"j AA199 SCrAgTb 2-12 . . . 3 SCrAgTb.. 6 M203 2-19 . . . AA284 12-7... 3 2 Sep 7 AA202 2-7 AA15 & 4 2 SCrAgTb 8 AA206 2-9 AA177 SCrAgTb lb-3 . . . 7 4 SCrAgTb 9 Do AA507 3c-22 . . 1 10 AA342 12-8. . . Do 4 11 A1058 fa M298 15-16.. ?, BrAgTb . . . 12 AA242 SCrAgTb 12-8. . . Bl 17 SCrAgTb ld-11.. 1 SCrAgTb . . TABLE 74. Cross 13. Miscellaneous matings of light-bellied agouti guinea-pig with non-agouti hybrid. No. 9 AgLb. cfNon-Ag. Lb Tb Non Remarks. Unclas- sified. }3392 G.p.... 3 1 SCrAgLb 1 3444 G.p.... 01 G.p.... JA1539 & fi 1 3R 2 3 02 G.p.... 3256 BrY AgLb G.p.... JA426R(Br) &..... A678 W & 7 ? 7 BrCrAgLb, 2 LBr 4 3220 BrAgLb G.p Do 2 2 BrAgLb 5 271 SAg(R) G.p. . . . M333Y jfc 4 4 SCrAgLb 6 3a G.p.... M34 Sep 16c-l . . ?, 2 Sep 7 241 SAg(R) ArF 2 .. M328 B-Y 42-17 . . 3 3 SCrAgLb 8 299 SAg(R) ArF 2 . . M156R iV 4 3 2 SAg(R) 4 W 9 12, 16 SAg(R) 21-1 . M34 Sep 16c-l . . 1 2 SCrAg, Sep TABLES. 135 TABLE 75. Cross 14. Miscellaneous matings of light-bellied with ticked-bellied agouti. No. 9AgLb. c?AgTb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 /20o, 5a G.p \3015,3014 G.p 20a,5a,3014 G.p AA173 14-1 . . . M82 9-7.... 49 9SAg(R) ArF,... JA412 R(Br) ^ 3 1 3R A1474 & 3 1 1 1 ?, AA199 SCrAgTb 2-12 . . . A1161 j 1 ! SCrAgTb . . AA508 3d-2 . . . 13 1 4 9 SCrAg, BY AgTb TABLE 76. Cross 15. Miscellaneous crosses involving the inheritance of agouti in rufescens hybrids. No. 9 Misc. cf Misc. Lb Tb Non Remarks. Unclas- sified. 1 AA171 R 14-1.. AA1161 AgTbg^ 1 R, Cr 2 M137R 9-1... AA286 AgTb 3c-4 . . 1 3 Do M29 Br 16-6 . . ? 4 MISlBrCrAgTb 15-15. C20 B G.p. . . 6 6 A1472 BrAgLb & 163B 4-toe.. ? 6 IA1562W ^ JAA83B jfc 4 1 8 \A1688AgLb 86-6. . f M27a AgLb 9-1 ... >393 B 4-toe . . 3 R ? 9 \M19 AgTb 2-16.. fM 175 AgLb Sa-3.. JA1040B t*ff 1 4 10 \AA242 SCrAgTb 12-8 . . [M 195 AgLb 9-7... JM116Sep 42-11. 1 2 11 \M203 AgTb 2-19. . fB 122 AgLb 86-2.. JC20 B G.p . . . ?, 2 2 1? \AA212 BrAgTb 2-6. . . FM92 AgLb 8a-l . . JA1513 5>j 1 2 1 BrAgTb 2R 13 \M106Y? 10c-2. [M85 R 16-5.. JAA1161 AgTb & 1 1 BrAgLb W 14 \M82 AgLb 9-7 . . [M56 AgLb 13-1 . . JA1 170 AgTb g^| 1 ?! 15 16 \M50AgTb 16-7. . AA28 W ^5 .... A1523 AgTb s>j A1513 AgTb s'j . . . . M83AgLb 9-7... 2 1 3 1 BrAgLb, SCrAg Tb, BrCrAg Tb R,2Cr 17 A1413B ,& Do 2 18 M84 R(Br) 16-5 . . A1161AgTb fc W 19 fA556AgLb ^ JA587W & [104B 4-toe.. 2 3 [A533Y & /A495AgLb - t V 1 3R 20 \A867B ^ /AA621SYALb 11-2.. >163 B 4-toe.. BW36 W BW... 1 SCrAgTb 21 \AA621 SYALb 724 SAg(R) lea ... 3 SSYAgTb 22 ?3 399SAg(R) ArF 2 .. 198W ArF 2 . . M224 BrAgLb 9-2... M291 B & 21 1 1 10 SAg(R), SAg Tb(R) W 24 29 9 W ArF 2 . . M2B j^g ?, 1 3 W 28 D125 W la-13 133 SAg(R) 24-1 . . 1 1 SAg(R),Sep(R). 26 D427W la-14. 133 SAg(R) 1 1 SAgTb(R), Sep 27 D86 W 76-3 . . 126 BWAg(R) 24-2 . . ?, 3 (R) 2 SAg(R), 3 Sep (R) 136 INHERITANCE IN GUINEA-PIGS. TABLE 77. Cross 16. Matings of dilute with albino of intense stock. Expectation: CaCd X C a C a = CdC a (all Dil). C d Cr X C a C a = C d C a + CrC a (1 Dil : 1 RE). C d C a X C a C a = C d C a + C a C a (1 Dil : 1 W). 16o. Females CdCd- No. 9 DU. cTW (intense stock). Int Dil RE W Remarks. 1 2 3 4 AA621S 3 YjAg 39-4 58 Seps Dil . . . BW36 W BW 1 SjCrjAgTb 2Sep s 2 Seps, Sepr-Cr* 3 S s CrAgTb 75 W BW . . ? Do 20 W BW... 3 Do . B42 W la-3 . 3 Total 9 ' 166. Females CdC a . No. 9 Dil. cf W (intense stock). Int Dil RE W Remarks. 1 2 3 4 5 6 7 8 9 10 11 12 12a 13 14 15 16 15 Sep Dil 17 Cr Dil 75 W BW.... B42 W la-3 . . . 4 1 3 2 Seps, 2 Sep 4 Sepe 30 Cr Dil Do 2 55 Cr 6 (Br) Dil Do 1 1 4 SjCrjAgTb 8Sep 6 3Sep 4 Sep 4 SjY^SsCrjAg Sep 4 SiCrsAgTb, 2 S, Y 4 Ag Seps-Oi Sep 3 , S 3 O 6 Ag Sepj, 2 Sep 4 Sep 4 , Sep 6 Seps /M42LBr 42-12... \M44Cr s 42-12... M42 LBr 42-12... J15W BW.... Do 8 3 AA600 LBr-Cr s 39-19 . . . Do 1 M357 S 4 CrsAg 42-1 .... B141 S 4 Y 4 Ag 39-23... 20 W BW.... Do 3 3 3 6 4 1 2 2 2 1 D43, D44 Sep 3 16o-3 . . . Do D45 Sj-Crj 16a-3 . . . . ..Do 1 D95 S 3 Y 4 Ag 166-9 . . . D95, M357 20 W BW.... Do ? D67 S 3 -Cr 6 16&-4 . . . ... Do M384Sep s 39-12... M442 BrCrjAgTb 39-12 . . 86 W BW.... .Do 3 ? D66 Ss-Crj 406-13 . . . .Do 1 1 32 Total 33 16c. Males CdC a . No. 9W. cfDil. Int Dil RE W Remarks. 1 2 3 4 5 6 132 W 4-toe.... 12oW 4-toe.... 5W BW A674 Sepe i 4 3 Sep 4 -Cr 6 , Sep 3 Sep8-Cr 2Sep 6 Ss-Crj, SjCrjAg Tb 2 S s Cr 6 AgTb Sep 4 2 SjYiAg, 3 S 4 - Y 4 , 2 S 4 , S M34 Sep 8 -Cr s 16c-l . . 3 Do 2 1 5 82 W BW /BW11W BW \BW15W BW 120, 21, 23, 29 W 22- Total B117 S 4 Cr s AgTb 39-14. . \ Do.. 2 3 / 13Cr,(Br) Dil 8 13 19 99 TABLES. 137 SIMILAR MATINQS FROM CROSSES 19, 27, AND 33, AND SUMMARY OF CROSS 16. No. Dilute. W (intense stock). Int Dil RE W 5 9 9 Dil (CdCr) W BW.... 9 10 3 o* d" Dil (CdCr) . ..Do 8 5 7 9 9 Dil (CdCa) .... . ..Do 5 7 1 cf Dil (CdCa) ... Do 4 ?, 16o 9 9 Dil (CdCd) .... W q 166 9 9 Dil (CdCa) .... . . .Do 33 3?, 16c d" cf DU (CdCa) . . . . Do ?,?, 19 Total 85 15 60 TABLE 78. Cross 17. Matings of intense from intense stock with albinos from dilute stock or from two dilute parents. Expectation: CC X C a C a = CC a (all Int). CC a X C a C a = CC a + C a C a (1 Int : 1 W). 17o. Male CC. No. 9W. cflnt. Int Dil RE W Remarks 1 M117 W 42-11 3013 B BB 2 2B 2 M327 W 42-17 Do ?, 2B 3 D 37 W Dil Do ?, 2B Total 6 17&. Female CC. No. 9 Int. cTW. Int Dil RE W Remarks. 1 22, 23, 33 Ag Misc 11 W Dil 11 11 Ag 2 CIS, C50 Ag Misc Do ft 6B 3 (C24 Ag Misc \ Do.. 4 Ag, 3B 4 \C34 B Misc 22 Br Misc / ... Do 4 4Br /S22B ,*,.. 6 }M313 W 42-16. . 4 4B 7 3223 B Misc J Do ? 2B 8 B232 B ld-21 M201 W 42-13 2 2B 9 B23AgTb ld-12 Do 1 AgTb Total 34 17c. MaleCC a . No. 9W. 9 Int. Int Dil RE W Remarks. 1 D37 W Dil 06 B BW ?, 2B 2 9 W Dil Do 1 fl B Total 3 9 138 INHERITANCE IN GUINEA-PIGS. TABLE 78 Continued. 17d. Female CC a . No. 9 Int. cfW. Int Dil RE W Remarks. 1 D48 Ag 176-1 11 W Misc 2 2 D49 Ag 176-1 Do 1 2 Br. 3 D224 Ag 176-1 Do 1 1 R 3o D224, D226 Ag 176-1 Do 3 2 Ag, 2B 4 B33 AgTb ld-18 M201 W 42-13 2 1 2 AgTb 5 B52 AgTb la-3 Do 2 2 AgTb 6 BllOAgTb la-1 Do 1 7 Bill AgTb la-1 Do 3 1 AgTb, 2 B 8 B128 AgTb la-1 Do 1 AgTb Total 13 10 SUMMARY OF CROSS 17. No. Intense (intense stock). White (dilute stock) . Int Dil RE W 17a cf cflnt CC.... W 6 176 9 9 Int CC... . Do 34 17c d"d"Int CC a . Do 3 2 17d 9 9 Int CC a Do 13 10 Total . . . 56 12 TABLE 79. Cross 18. Intense guinea-pigs, each of which had a dilute parent known to transmit albinism, mated with albinos or red-eyes to test whether the same intense animal can transmit both dilution and albinism. Expectation: CCa X C ra Cr a = CC ra + CaC ra (1 Int : 1 Dil). CC a X CraCra = CC ra + C a Cr a (1 Int : 1 RE or W). 18o. Male CC d by test. No. 9 Red-eye. cflnt. Int Dil RE W Remarks. 1 2 3 4 5 6 7 8 9 10 515SAg(R) ArF 2 .. 774 SAg (R) ArF 2 .. D 10 R (Br) 35-1 . . . Do 3 1 3 3 Ag.SCrAg, 2 S 3 YAg Ag 3 Ag,2 S!iY4Ag, S 4 Cr 6 Ag 4 B,2 Si-Crs, S 4 -Y 4 3 Ag.SsY^Ag, S 4 Y4Ag S 6 Cr 6 Ag, 2 SiYiAg 2 S 4 Cr5Ag, S 6 Cr, Ag 3 Ag,2 S 3 Y 4 Ag, S 6 Y 4 Ag 2 B,S 3 Cr 5 Ag 3 Ag,BiY 4 AgTb, S 3 Cr B Ag Ag.SsY^g, SjY^g 4 S 4 Cr5Ag, SCr 6 Ag 2B^ 515, 774 SAg (R) ArF 2 .. Do 3 4 3 3 2 3 1 2 3 3 8 3 1 2 7 514Sep(R) ArF 2 .. ... .Do 281 SAg (R) ArF 2 .. 741 SAg (R) ArF 2 .. D 30 R (Br) 36-1 . . . Do 241 SAg (R) ArF 2 .. 413 SAg (R) ArF 2 .. AA508AgTb 3d-2... Do 759 SAg (R) ArF 2 .. Do M430 SAg (R) 18r-4 . Total 3 males Do 25 30 TABLES. TABLE 79 Continued. 139 186. Female CCd by test. No. 9 Int. C? Red-eye or W. Int Dil RE W Remarks. 1 2 3 M292 B & . . . . M353 B A D 18 W 166-3 . . Do 2 1 2 5 3 1 1 5 AgTb, B,S 6 Cr 6 AgTb, 2Sep s AgTb,Sep 6 (Cross 20) D6a R (Br) 35-1 . . Total 3 females 724 SAg (R) lea 18c. Male CC a by test. No. 9 Red-eye or W. cflnt. Int 5 91 Dil RE 3 8 W 1 Remarks. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 271 SAg (R) ArF 2 .. 278SAg(R) ArF 2 .. M224 BrAg 406-12 . Do 5Ag,2S4Ag(R),S 7 Ag (R) 2Ag,5SAg(R) 4Ag,S&Ag(R),SAg(R) SiAgTb (R) 2 Ag, 3 B,2 SAg (R) AgTb.SgAgTb (R) Seps (R) 2 AgTb,Sep 4 (R) 4B,Sep 6 (R),2Sep(R) 2 Ag,Se P4 (R) Ag, B 2 Br, Ag 5B, Ag 11 B 2 Ag, 8 B 716 SAg (R) ArF 2 .. Do 4 3 233 SAg (R) ArF 2 .. 773 SAg (R) ArF 2 . . M156R(B) ^ Do 5 2 3 1 1 236Sep(R) ArF 2 .. 264 Sep (R) ArF 2 AA433aAgTb 36-4... Do 1 2 2 1 485Sep(R) ArF 2 .. 235 Sep (R) ArF 2 .. 693 W ArF 2 .. D7 R (Br) 35-1 . . . D13 R (Br) 35-1 . . . Do 4 2 a 3 1 1 2 3 9 13 7 41 D42 W 166-1 . AA578 W 3c-18 . 3 9 9 W ArF 2 .. 4 9 9 W Misc. . 3 9 9 W ArF 2 .. Total, 8 males Do Do M291 B ^j 3 6 M339 B 40o-14 . M2B i>g 11 10 57 20 140 INHERITANCE IN GUINEA-PIGS. TABLE 80. Cross 19. Matings of dilute from cross 18o or 43, with albino. Expectation: C d C r X C a C a = C d C a + CrC a (1 Dil : 1 RE) (1-6). C d C a X C a C a = C d C a + C a C a (1 Dil : 1 W) (7-13). No. 9 Dil (or W). cfW (or Dil). Int Dil 1 1 2 RE W Remarks. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 D72S 3 Y4Ag 18a-5.. D63 S 4 Y4Ag 43-2 . . . BW36W BW ... 2 3 2 ?: Sr-Y 4 ,StAg(R),Sep t (R) Sep 4 , 2 SjAg (R), Sep 4 (R) SaCrsAg, SrOi, SsAg(R) Sepa (R) SsAg(R), Sep(R) S 4 Ci5Ag, Sep 4 2 S 5 Cr6Ag, Sj-Cre, S^g (R) S,Y 4 Ag 2 S 4 CrsAg, Sep 4 S 4 O(Ag 3 S 4 CrftAg SsY^g, S 4 -Cr 6 3 S.Y.Ag Sepe Do D121W 18o-9.. 157 W 22-3... D71 SsOiAg 18o-5. . Do D148W lc-8... D239W 18c-6.. D240 SjY^g 18o-9 . . 2 Do 3 1 DGlSsY^Ag 43-2... D241S 6 CrsAg 18o-9.. /D69S 4 Cr&Ag 18o-5.. \M425SYAg 43-1... S772W ArF 2 ... 256 W ArF 2 ... BW36W BW.... 1 1 1 BW50W BW.... JBW36 W BW. . . . D70S 4 Y 4 Ag 18a-5.. 1 3 1 Do 3 3 1 S781W ArF 2 ... Do ? D69S 4 Cr 6 Ag 18o-5.. D206S 4 -Y 4 18a-4.. 2 females C d Cr BW36W BW.... T BW50W BW.... 1 2 7 ?, 5 5 3 4 2 males C d Cr 3 females C d C a 1 male C d C a 6 2 females (?) fi TABLE 81. Cross 20. Matings of pure lea male 724 CrC r . Expectation: CC X CrCr = CCr (all Int) (1). CC d X Q-Cr = CCr + C d Cr (1 Int : 1 Dil) (2-3). C d C d X CrCr = C d Cr (aU Dil) (4-5). C d C a X CrCr = C d Cr + CrC a (1 DU : 1 RE) (6). No. 9 Misc. cf Red-eye lea. Int Dil RE W Remarks. 1 2 3 4 5 6 5 9 9 B Trior4-toe D6aR(Br) 35-1 . . . 724SAg(R) lea... q 9Ag 2 Ag, SzY 4 Ag Ag 2S 2 Y 4 Ag SSjY^g BiCr&Ag, S 2 D196Bi 26-2.. 1 2 S 6 (R), S,(R), 2 S 7 (R). S 8 (R) Sj 3 G30 W St Do 1 3 S 6 , S4(R),S 6 (R),87 4 D115Bi-Cr 6 26-1... BW50 BW . 3 (R) 3 S 3 (R) 5 D197Bi 26-2... JBW46 BW 3 ? 3 S 4 , 2 S 6 (R) 6 7 D198B t 26-2... BW43 W BW . DllSBiOsAgTb 26-1 ?, S 4 , Ss-CrsAgTb 8 9 10 11 482 W ArF 2 ... D42 W 166-1 . . BW56.57 BW.... D122 SrCrsAg 26-4 D123Si-Cr 6 26-4.. D55S 5 -Cr 5 26-2.. D114Sep 6 26-1.. BW50 BW 3 4 4 1 3 2 ? S 6 , 2 S 6 -Y 4 85, Sc, S 4 , Se Cre 2 S 4 , 2 S4-Y 4 12 D195 Sep 4 26-2 Do 1 1 S< TABLE 89. Cross 28. Matings of pure Arequipa male 1007 C d C d with black guinea-pigs. Expectation: CC X C d C d = CC d (all Int). CC a X C d C d = CC d + C d C ft (1 Int : 1 Dil). No. 9 Intense. cf Dilute (Arequipa). Int Dil RE W Remarks. 1 2 9 9 B 4-toe ^007 SYAg .... 4 3Ag, B 2 49 9B BW ... Do 10 5Ag, 5B 3 1442 B BW Do 3 2 2 Ag, B, SCrAg, Sepi '1007 SYAg CdCd from 1001 BRAg CCd and 1002 SCrAg CdCr pure Arequipa stock. TABLES. 143 TABLE 90. Cross 29. Matings of intense FI (cross 28) with of 1007. Expectation: CC d X C d C d = CC d + C d C d (1 Int : 1 Dil). No. 9 Intense ArFi. c? Dilute (Arequipa). Int Dil RE W Remarks. 1 SA4 Ag 28-3 .... 4007 SYAg 3 2Ag, R 2 SA8 Ag 28-2 ..Do. . 9 2 AR, R, Bi-Yj, Y s (SA4, SA8 \ T^ 3 > Do 3 2 2 Ag, B, BiYsAg, Y 3 \SA10 B 28-2 / J 1007 SYAg C d C d from 1001 BRAg CC d and 1002 SCrAg C d Cr pure Arequipa stock. TABLE 91. Cross SO. Mating of dilute FI (cross 28) with c? 1007. Expectation: C d C a X C d C d + C d C d + C d C a (aU Dil). No. 9 Dilute ArFi. 9 Intense (Arequipa). Int Dil RE W Remarks. 1 SA3 Sep 3 28-3 r 1007 SYAg 2 BiYAg, Y! J 1007 SYAg C d C d from 1001 BRAg CC d and 1002 SCrAg C d C r pure Arequipa stock. TABLE 92. Cross 31. F s from intense FI Arequipa (cross 28). Expectation: CC d X CC d = CC + 2 CC d + C d C d (3 Int : 1 Dil). No. 9 Intense ArFi. cT Intense ArFi. Int Dil RE W Remarks. f SAG B 28-2 1 >SA2 B 28-2 17 3 17 B, 2 B-Ys, Bj-Y \SA1 1 B 28-2 2 SA13 B 28-3 .... SA7Ag 28-2. 1 B fS' --. = . 3 fSASAg 28-2 \ Do . . 3 Ag, 2B 4 \SA10 B 28-2 SASAg 28-2 / Do 1 2 Ag, B 2 YAg, Br-Ys 5 SA4 Ag 28-3 Do 1 91 Ag, 2 B!YAg Total ?3 7 TABLE 93. Cross 82. F 2 from dilute Fj X intense Fi (cross 28). Expectation: C d C a X CC d = CC d + CC a + C d C d + C d C a (2 Int : 2 Dil). No. 9 Dil ArFL d"Int ArFi. Int Dil RE W Remarks. 1 SA3 Sepa 28-3 SA7 Ag 28-2 1 1 B, B 2 Y 3 Ag 2 Do SA12 Ag 28-3 1 1 B, 84 144 INHERITANCE IN GUINEA-PIGS. TABLE 94. Cross 33. Mating of dilute FI Arequipa with albino. Expectation: CaC a X C a C a = CaC a + C a C a (1 Dil: 1 W). No. 9 Dil ArF,. d" White. Int Dil RE W Remarks. 1 SA3 Sepa 28-3 ... .... 75 W BW 1 TABLE 95. Cross 34. Matings of intense F t Arequipa with albinos. Expectation: CC d X C a C a = CC a + C d C a (1 Int : 1 Dil). No. 9 Int ArF! (or W). cfW(orlntArFi). Int Dil RE W Remarks. 1 SA4 Ag 28-3 . . M313W 42-16. ? 2 Y4 2 SA10, 11, 13 B 28-2, 3 . Do . . . 4 R 3 B, R, 2 S s , 4 Ss-Cn. 2 Y 4 3 149 W 22-2 .... SA26 Ag 28-1. . . 8 ?, 2 Ag, B, SaYiAg, SjCrjAg 4 161 W 24-1 .... Do 3 1 2 Ag, B, 82 5 349 W ArF 2 Do 1 1 Ag, 83 6 149, 349 W Do 2 6 Ag, B.SsCrsAg, 2 S 6 YtAg, 2 SsCrsAg, S 3 -Crj Total . . . 13 ?n TABLE 96. Cross 85. Mating of cream of dilute selection stock with a red stock free from albinism or dilution. Expectation: CC X CdC a = CC d + CC a (all Int). No. 9 Intense. cf Dilute. Int Dil RE W Remarks. 1 499 R(Br) Misc. 00 Cr 6 (Br) Dil 12 11 R(Br), Y 2 (Br) TABLE 97. Cross 36. FI (cross 35) mated with father. Expectation: CCd X CdC a = CC d + CC a + CdCd + C d C a (2 Int : 2 Dil) (3). CC a X C d C a = CC d + CC a + C d C a + C a C a (2 Int : 1 Dil : 1 W) (1-2). No. 9 Intense FI. cf Dilute. Int Dil RE W Remarks. 1 2 Dl R(Br) 35-1 D2 R(Br) 35-1 OOCr 6 (Br)Dil... Do 4 6 1 2 1 4 R(Br), 2 Y 4 (Br), 4 Cr 8 (Br) Cr 6 (Br) 3 D6 R(Br) 35-1 Do s 4 3 R(Br), Y 2 (Br), 3 Cr B (Br) 4 D136 R(Br) 36-1 Do 9 1 2 R(Br}, Cr 6 (Br) - TABLES. TABLE 98. 145 Cross 37. Matings of dilute with dilute in the dilute-selection stock. Expectation: C d C d X C d C d = C d C d (all Dil). C d C d X C d C a = C d C d + C d C a (all Dil) (1). C d C d X C d C a = C d C d + 2 C d C a + C a C a (3 Dil : 1 W) (2-11). No. 9 Dilute. c? Dilute. Int Dil RE W Remarks. 1 D301 Y 4 Dil .... D292Y 4 Dil... 3 Y 3 , 2 Cr 6 2 Do D298 Cr 6 Dil 2 Y 8 , Crs 3 D300 Cr 6 Dil .... Do 3 1 Y 3 , 2 Cr 6 4 D299Cr Dil Do 5 Y 4 , 3 Cr 6 , Cr 5 6 D289Cr Dil.... D291Cr 6 Dil D290Cr Dil... Do 6 1 2 Y 3 , 4 Cr 6 , Cr 7 8 /D260Y 3 (Br) Dil.... \D262, D263 Cr 6 (Br) Dil. ... D263Cr g (Br) Dil.... JD261Cr 6 (Br)Dil... Do ... 4 3 1 2 Y 8 (Br),3Cr 6 (Br) Y 3 (Br),2Cr 6 Br 9 D262Cr 6 (Br) Dil Do . 3 2 Y 3 (Br), Cr B (Br), 10 [D265 Cr 6 (Br) 37-7.. 1 Do. . 4 Cr 6 (Br) Y 3 (Br),3Cr 6 (Br) 11 \D266 Cre(Br) 37-7 . . D265 37-7 . . / Do . . 2 1 Y 3 (Br), Cr 6 (Br) Total (excluding 1) . . 32 10 TABLE 99. Cross 38. Matings of dilute with albino in the dilute-selection stock. Expectation: C d C d X C d C a = C d C a (all Dil). C d C a X C a C a = C d C a + C a C a (1 Dil : 1 W). No. 9 White (or Dilute). d" Dilute (or W). Int Dil RE W Remarks. 1 38a. D293 W Dil D292 Y 4 Dil 4 4Cr s 2 D294 W Dil Do 2 2Cr 6 3 D264 Y 3 (Br) 37-7 11 W Dil ?, 2 Cr(Br) 2C d C d 8 1 386. D293 W Dil D290 Cr Dil 2 2Cr 4 2 D302 W Dil D298 Cr 5 Dil 1 Cr 6 3 D303 W Dil . .Do 2 2Cr 5 4 D 75 W Dil D267 Cr 6 (Br) 37-7 4 5 Do D274 Cr 6 (Br) Dil ?, 1 2 Cr 6 (Br) fD276 W Dil 6 >D261 Cr 6 (Br) Dil 3 2 3 Cr(Br) \D272 W Dil . ... 5 C d C a . 10 7 146 INHERITANCE IN GUINEA-PIGS. TABLE 100. Cross 89. All matings of intense with intense which have given dilute young, except those given in cross 31. Expectation: CCd X CC d = CC + 2 CC d + C d Cd (3 Int : 1 Dil) (1-7). CC d X CC a = CC + CC d + CC a + C d C a (3 Int : 1 Dil) (8-33). 3, 9, 11, 24, 26, 30, and 32 not wholly certain. No. 9 Intense. d" Intense. Int Dil RE W Remarks. 1 M353 B ^j SA26 Ag 28-1 . . . 3 1 Ag, B, R, Sj-Yj 2 A780 AgTb fa A781 AgTb fa 7 2 4 AgTb, 3 R, BrYAg 3 B58 AgTb ld-15 . . B 118 AgTb ld-6... 9 2 Tb, Y 5 AgTb, 3 BrAgTb, 4 M25 AgLb 9-1 M91 AgLb 8-4 6 4 B, SjYjAgTb, St 3 Ag, 3AgTb, 2 SYAg 5 AA588 Ag 39-4 . . . .Do 2 1 Lb, 2 S 3 Y 2 AgLb 2 Ag, S 3 YsAgTb 6 f M25, M27a Ag 9-1 .... > .Do 3 1 [Ag, 2 AgTb, BiYiAg 7 \B139 Ag 39-23 . . M177B lc-2... /" .Do 4 1 \ Tb Ag, 3 AgTb, SYAgTb g M168B Ji . . Do 4 1 2 Ag, 2 AgTb, SCrgAg 9 M169, M171 B 3^ Do 3 3 3 AgTb, 2 BrOAgTb, 10 B68 AgTb la-1 B 118 AgTb ld-6 6 1 BrYAg 4 AgTb. 2 B, SjCr 5 11 A443 B t A469 AgTb $ 1 AgTb LBr 12 M90 Br ^ M 189 AgTb 39-30. . 3 BrOsAgTb, S 6 , Cr 6 13 /M90Br & Do 6 2 3 AgTb, 3 B, S 5 Cr 6 14 \M114B 16-7... A1117B 3*1 1357 B BW.... 2 1 AgTb, S 5 AgTb, B, S 4 Cr 6 AgTb 15 A1566AgTb j% . . A1050 AgTb 3*2 4 2 3 AgTb, BrAgTb, SCr 16 Do AA15AgTb sV 2 1 AgTb, S 4 AgTb, B, SCrAgTb 17 A529 BrAgTb 3*3 Do 5 2 3 AgTb, B, BrAgTh, 18 AA202 AgTb 405-8 . . . . .Do 2 2 SCrAgTb, BrCrAg Tb 2 AgTb, 2 SCrAgTb 19 M177B lc-2. . . AA235 AgTb 406-7 . . 4 1 AgTb, 2 B, BrAgTb, 20 M 102 AgLb 66-1... M2B &. 4 1 LBr-Cr s 2 Ag, 2 B, SjCr&Ag 21 3392 AgLb Misc . . A1539 B -fa 1 SCrAg 22 3392, 3444 Ag Misc . . Do ... 3 1 2 Ag, B, SCrAg 23 3392 Ag Misc . . B5 AgTb ld-16 3 1 3 Ag, S 4 Y4Ag 24 20a Ag Misc . . A1474 AgTb A 3 1 3 Ag, SCrAg 25 A1310Ag jV A1449 AgTb s\ 2 1 AgTb, B, Cr 26 M203 AgTb 2-19 . . . AA284 AgTb 39-18 . . 4 1 3 AgTb, B, S 27 M82 Ag 9-7 A1161 AgTb sV 1 1 AgTb, SCrAgTb 28 AA171 R 3*5 Do 2 1 AgTb, R, Cr 29 M183B ^ AA299 AgTb 40-6 . . . 3 1 3 AgTb, SCrAgTb 30 20 B Misc . . A412R(Br) &. . . 3 1 2 AgTb, B, SCrAgTb 31 M7B & M133Ag 8-4 4 ? 2 Ag, 2 B, SCrAg. S 6 32 A1420B h- A811 Br sV 1 1 Br, LBr 33 A385 B jf 12845 B 4-toe... 4 1 4B, S Total 109 H7 1 Ezcess of dilutes expected because the presence of at least one dilute young is used aa a criterion for admission to the table. TABLES. 147 TABLE 101. Cross 40. All coatings of intense with dilute which have given dilute or albino young, except those of crosses 28, 29, 32 and 36. Expectation: CC d X C d C d = CC d + C d C d (1 Int : 1 Dil), 1. CC a X C d C d = CC d + C d C a (Int : 1 Dil), 2-5 (5?). CC d X C d C a = CC d + CC a + C d C d +C d C a (1 Int : 1 Dil), 6-19 (9, 10, 12, 14, 15, 16, 17?). No 9 Int (or Dil) . cfDil (or Int). Int Dil RE W Remarks. 1 40a. B139 Ag 39-23 . . M328Bj-Y 4 42-17. ?, 3 2 AgTb, 3 SjY t 2 AA606 SjY.AgTb 40o-8 . . AA573 Br AgTb 40o-7 . 1 1 Ag AgTb, S t Cr 6 Ag 3 4 9 9 B Misc. . . AA241 SYAgTb 40o-6. 5 3 Tb AgTb. 4 B, S 4 169 B 4-toe... A656Br-Y | 1 3 CrjAgTb. Sr- Cr, LBr B, 3S-Cr 5 AA497 SYAgTb 39-7. . . AA284AgTb 39-18. ?, 1 AgTb, R, LBr 6 AA206 AgTb 39-2 . . . AA177 SCrAgTb 41-4. . 3 4 3 AgTb, SYAg 7 AA213AgTb 39-15.. AA253 S 8 Cr 6 AgTb 406-8. 4 5 Tb, 2 S,Y,Ag Tb, S,Cr B Ag Tb 3 AgTb, BrAg g AA217 AgTb 406-8 . . Do 7 4 Tb, SjYjAgTb, BrYsAgTb, S s CrsAgTb. S 6 CrjAgTb, Br Cr^AgTb 7 AgTb SjYtAg g AA613AgTb 40o-7.. Do 3 Tb,2S 3 Y^g Tb, SCrAgTb S^YaAgTb SsCrj 10 11 12 3o AgLb Misc. . . M261Sep4 41-2... M282 Ag 15-12 . . M34 Sepr-Cr t 16e-l . AA299AgTb 40o-6. M 1 16 Sep s -Y 4 42-1 1 . 3 2 1 1 AgTb, S 4 Cr s AgTb 2Sr-Cr 8, 2 AgLb, B, S- 13 30 Br Misc. . . M34 Sep-Crj 16c-l . 1 ? Cr, R, B 2 , S 14 M99 B 42-13? . Do .... ?, 1 2 B, S-Cr e 15 M101B, M99B 42-13?. Do 3 1 B, 2 R, 87 16 M99 B 42-13? . A674 Sep 6 J 1 Sr-Cr? 17 M155 B ^ Do . 3 2 S, S 7 18 19 M82Ag 9-7 SA13B 28-3... Total M116Sep 5 -Y 4 42-11. M306S7-Cr 7 42-15. 2 S6 3 2 441 2 AgTb, 8Y,Ag Tb, BrYAg Tb, SCrAgTb Ss-Crj, Sr-Crr 'Excess of dilutes expected. 148 INHERITANCE IN GUINEA-PIGS. TABLE 101 Continued. Cross 40 Continued. No. 9Int (or Dil). d*Dil (or Int). Int Dil RE W Remarks. 1 2 406. 1 A1659 B *& A 1523 AgTb ife AA253 S 5 Cr s AgTb 406-8. AA199 SCrAgTb 39-15 g 2 1 2 SCrAgTb 2 AgTb, SCrAg 3 B132AgTb ld-3... M293 Y 4 42-14 . 1 1 Tb AgTb, SeCr 6 Ag 4 5 M442 BrCrsAgTb 39-12 . . M44Cr 6 42-12.. AA573BrAgTb 40a-7. AA197AgTb 2-10.. 2 1 1 Tb BrCr 6 AgTb 2 AgTb, SCrAg 6 7 8 M181 BrCr 6 AgTb 41-6. . . AA203 BrCrAgTb 39-17. . A1273SCrAgTb 3"$ AA573 BrAgTb 40o-7 . AA16 AgTb g>5 . . . . Do 8 2 8 2 2 3 2 1 3 Tb 3 BrAgTb, Br Cr B AgTb, Br Cr 6 AgTb 2 AgTb, SCrAg Tb, LBr AgTb, 2 B, SCr 9 10 AA176 AgTb 41-4 . . . AA175 AgTb 41^1 . . . AA177 S 6 Cr 6 AgTb 41-4.. Do 5 1 1 1 AgTb, S B Cr B AgTb, Sj-Y 4 4 AgTb, B SCr AgTb 11 AA671 AgTb 40o-7 . . AA253 S 6 Cr 6 AgTb 406-8. 1 12 13 M46 BrCreAg 44-1 . . . S443Sep ArF 2 ... Total A1170AgTb sSf.... M156R & 2 4 ?4 1 15 2 1 2 I? 2 Ag, BrAg 4B,S r -Cr B Expectation: CC a X CdCa = CCd + CC a + CdC a + C a C a (2 int : 1 dil : 1W). 2 Excess of dilutes and albinos expected. TABLE 102. Cross 41 - All matings of intense with albino which have given dilute young, except those given in crosses 18 and 34. Expectation: CC d X C a C a = CC a + C d C a (1 Int : 1 Dil). No. 9 Int (or W). <7W (or Int). Int Dil 2 RE W Remarks. 1 2 3 4 5 6 7 Al 146 AgTb jV ... M102 AgLb 66-1 . . B139Ag 39-23. A1227W V--- A1309W .... AA28W &.. A504 W t"s . . . S, S-Y B, S 4 AgTb, S 3 Cr 6 AgTb 2 AgTb, S 6 Cr 6 AgTb BrAgTb, B. R, R(Br), Cr(Br) Ag, BrAg.AgTb, R, SCrAgTb, BrCrsAgTb, 2 Cr B, S-Cr A462W &... 20 W BW . A781AgTb &... A1513AgTb ^... ... Do 1 1 2 4 4 1 13 1 1 1 1 4 1 II 1 131 W 4- toe. . Total A412R(Br) ^. . . Excess of dilutes expected though not found. TABLES. 149 TABLE 103. Cross 4&> All matings of dilute with dilute, except those of crosses 30 and 37. Expectation: C d C d X CdC a = C d Cd + C d C a (all Dil) (M328, AA242, M394, C d C d ). C d C a X C d C a = C d C d + 2 C d C a + C a C a (3 Dil : 1 W). No. 9 Dilute. d" Dilute. Int Dil RE W Remarks. 1 B141 S 4 Y4Ag 39-23.. M328 Br-Y 4 42-17 . . 4 3 ^YaAgTb 84 2 AA242 SjYgAgTb 40o-6 . . B117 S 4 CrsAgTb 39-14 . . 1 CrjAg S 6 Cr 8 AgTb 3 AA244 Sep 4 39-15 . . Do 1 SgCrsAgTb 4 D44 Sep 3 16o-3 . . Do 1 ? S 4 Cr&AgTb 5 D43 Seps 16a-3 . . D94 S 3 Y*Ag 165-9 . . 2 S 3 CrsAg 83 5a D45 Sepr-Crs 16a-3 . . Do ? 1 S 2 SsCr^Ag 6 D26 S 5 Cr&AgTb 16a-4 . . D33 SsCr^AgTb 166-^ . . 1 7 8 /D110Sep 6 16c-3.. \D107 Sep 4 166-1 . . D215 Seps 16o-2.. JD 106 Sep4 166-1 . . Do 5 ? 1 5S 3 Bi Si 9 10 M181 BrCrjAgTb 41-6. . . 3520 Y 4 (Br) Dil AA253 S 8 Cr 6 AgTb 406-8 . . A674 Sep 8 \ 1 6 1 1 SY4AgTb S^YiAe SsCrs 11 3417 Cr Dil . . . .... Do s 9 Ag, BrCreAg, Y4(Br),2Cr (Br) SK YJ Y^ Cr 12 /3417 Cr 6 Dil .... } Do. . 4 S-Cr, LBr, Y, 13 \3462 Cr 6 (Br) Dil .... O6 LBr-Cr Misc. / .Do 19 4 1 Cr 6 2B S-Cr 87- 14 M127Cr 6 42-13.. Do 1 Cr 7 ,Cr 6 (Br), Cr & Y 4 15 M44Cr 6 42-12.. Do 1 1 &j-C^ 16 M164Cr 7 44-6... Do 9 17 M126 Cre(Br) 42-13. . Do 1 1 B Y 4 18 19 M164Cr 7 44-6... M44 Cr B 42-12 . . M306 Sep 7 -Cr 7 42-15 . . Do 2 3 1 S 7 -Cr 6 , Cr 6 Se-Cre Cr 7 , Cr 20 M296 SCrAgTb 39-27 . . M116Sep 6 -Y 4 42-11.. 3 SY2AgTb 2S 21 22 MSlOSepT-Crr 40a-16. M336 Sep 40a-17 . M335Sep 6 40a-17. Do 1 3 1 1 Y 3 AgTb S 6 23 M336 Sep 6 M34 Sepe-Crs 16c-l . . 1 S 4 -Ys 24 M394 Sep 4 42-22 . . Do 3 S 4 -Y 4 Yj Crj 25 /M336,Sep 6 40o-17. 1 Do . . 4 26 \M394Sep4 42-22.. M393 Sep 6 42-22 / Do 1 27 D278 Cr 5 (Br) 386-5 . . Total D138 Cr 6 (Br) 36-1 ... 2 1 60 1 19 Cr 6 (Br) C d C d X C d C a 10 CdCa X CdCa 50 19 'Recorded from a mixed pen before the study of dilution was begun, probably an error. TABLE 104. Cross 43. All matings of dilute with red-eye, except those of crosses 20 and 26. Expectation: C d C d X C r C a = C d C r + C d C a (all Dil) (1-2). C d C a X CrC a = C d Cr + C d C a + CrC a + C a C a (2 Dil : 1 RE: 1W) (3-4). No. 9 Red-eye. cf Dilute. Int Dil RE W Remarks. 1 241 SAg(R) ArF 2 .. M328B2-Y 4 42-17.. 3 3 SCrAg 2 271 SAg(R) ArF 2 . . M333Y 2 42-11.. 4 2 S 6 Y4Ag 2 S 4 Y4Ag 3 236 Sep(R) ArF 2 .. M331 BrCrAg 42-10 . . 1 4 D194 Sep4 Do 5 1 4 \3770 4-toe 3769 4-toe I" 3987 4-toe a 5 3770 4-toe Do ? Total 10 3 TABLE 107. Cross Ifi. Rough A (tricolor) X rough A (tricolor) ; one or both of parents of each, rough C or D. Rrss X Rrss = 3 Rss + rrss (3 A : 1 Sm) (1-8). or Rrss X RRss = Rss (all A) (9-12?). No. 9 Rough A. c? Rough A. A B C D E Sm 1 4018 Tri 3775 Tri 1 2 3941 Tri 3940 Tri 3 3 3943 Tri Do 1 4 /R65 54-17 \3943 Tri | Do 7 1 5 R65 54-17 Do ?, ?, 6 R171 47-3 Do . 2 2 7 R278 52-14 R248 52-10 1 1 8 R357 Red 52-8 Do 1 1 1 9 R65 54-17 R197 52-14 2 10 R171 47-3 Do 3 11 R194 54-1 . . .Do 2 12 R196 52-14 Do 3 Total 1 to 8 17 ' 7 Total 9 to 12 . . 10 V TABLES. 151 TABLE 108. Cross 47. Rough A X rough C (tri); all mothers of tricolor stock except R175 4-toe. Rrss X RrSs = 3 Rss + 3 RSs + 2rr (3 A : 3 C : 2 Sm). No. 9 Rough A. d" Rough C. A B C D E Sm 1 R21 46-2.... R52 56-1 1 2 2 R23 46-2 .... Do 1 1 3 R42 50-1 .... Do 1 1 1 4 R21 46-2.... R99 56-1 1 2 5 R23 46-2 .... Do 1 1 6 R42 50-1.... Do 3 1 2 4 7 R175 49-1 .... Do 3 1 Total . . . 10 1 5 1 10 TABLE 109. Cross 48. Rough A (Tri) X rough E (Tri). Rrss X RRSS = RSs (all C). No. 9 Rough A. cf Rough E. A B C D E Sm 1 R42 50-1 4003 Tri 2 1 TABLE 110. Cross 49- Rough A (4-toe) X smooth (4-toe). Rrss X rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Smooth. cf Rough A. A B C D E Sm 1 499 4-toe 3922 4-toe 18 13 2 599 4-toe 3609 4-toe . 10 1 19 Total .... ?8 1 3?, TABLE 111. Cross 50. Rough A, B (tri) X smooth (4-toe etc.). Rough A, B with one or both of parents paitial rough. Expectation as in cross 49. No. 9 Smooth (or rough B). c? Rough A (or smooth) . A B C D E Sm 1 799 Sm 4-toe 3775 A Tri 13 14 2 R121 Sm 50-1 Do 3 3 R62 Sm 50-1 R22 A 46-2 3 3 4 IR163 B 52-13 99 Sm 4-toe 3 Total 19 90 'R163 may be RR. 152 INHERITANCE IN GUINEA-PIGS. TABLE 112. Cross 51 . Rough A X smooth (tri) ; smooth with one or both parents partial rough. Rrss X rrSS = RrSs + rrSs (1 C : 1 Sm). Rrss X rrSs = Rrss+ RrSs + 2rr(lA:lC:2 Sm). Rrss X rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Smooth. cf Rough A. A B C D E Sm 1 R13 52-1 R22 46-2 1 3 4 2 R123 54-3 R76 4-toe 2 4 3 R124 54-3 Do 1 2 4 R133 47-2 Do 2 5 R124, R133 Do 2 8 6 R142 54-3 Do 1 1 Total 6 3 3 19 TABLE 113. Cross 52. Rough 1 C, D (tri) X rough C (tri). RrSa X RrSs = 3 Rss + 6 RSs + 3 RSS + 4rr(3A:6C:3E:4 Sm). No. 9 Rough C, D. d"Rough C. A B C C E Sm Remarks. 1 3013 Tri 3780 Tri 1 3 1 1 ? 2 3246 Tri Do 1 ? 3 Rll 52-1 Do 1 2 3 4 R54 52-1 .... . . .Do 1 1 5 3245 Tri 4019 Tri . ... 2 1 6 3939 Tri R58 52-5 . 4 1 1 Red-A 7 3809 Tri Do 2 1 Red-E, Red-Sm 8 3246 Tri Do ?, 1 1 Red-A g 3724 D Tri Do 1 1 10 3724 3246 Tri Do 1 4 1 11 R51 56-1 R56 56-1 . ... ? 1 2 2 1 s 12 13 R57 56-1 R51 R57 56-1 Do . . Do 1 1 1 2 14 R98 56-1 Do . . 5 ? 3 1 15 R103 56-1 Do 1 1 5 1 1 Red-C Total 18 A 19 7 }?, 17 1 AII rough C except 3724. TABLE 114. Cross 53. Rough C, D (tri) X rough E (tri). RrSs X RRSS= RSs + RSS (1 C: 1 E). or RRSsX RrSS = RSs + RSS (1 C : 1 E). or RrSa X RrSS = 3 RSs + 3 RSS + 2 rr (3 C :3 E : 2 Sm). No. 9 Rough C, D. d" Rough E. A B C D E Sm 1 R6 D 54-15 .... 4003 Tri 1 ?, 2 R88 C 52-1 . . . R200 Red 52-7 2 3 R286 C 52-6 . R280 52-14 .... 3 4 R222 C 52-12 .... Do 1 ?, 1 Total 4 1 6 1 TABLES. 153 TABLE 115. Cross 54. Rough C, D (tri) X smooth (4-toe, etc.). RrSs X rrss - Rrss + RrSs + 2rr(lA:lC:2 Sm). No. 9 Smooth (or rough C,D). cf Rough C,D (or smooth). A B C D E Sm 1 399 Sm 4-toe . 3780 C Tri 1 1 8 2 R62 Sm 60-1 Do ? 2 3 599 Sm 4-toe R12 C 52-1 4 12 17 4 399 Sm 4-toe R26 D 54-15 8 3 fi 14 5 499 Sm 4-toe R52 C 56-1 2 2 6 6 599 Sm 4-toe R102 C 56-1 4 5 3 9 7 B31 Sm ld-9 R105 C 48-1 2 g M253Sm la-10 R106 C 48-1 2 9 M255Sm la-10 Do 1 10 M380 Sm ^j Do 1 11 131, M253, M255 Sm Do 1 3 12 M384 Sep-Sm 1&-9 R99 C 56-1 . 2 2 13 R62 Sm 50-1 R112C 52-11 3 14 65 Sm 4-toe . . . Do 1 1 1 15 3246 C Tri 2967 Sm BB 5 1 2 3 16 3809 C Tri Do ... 1 3 17 3724 D Tri . . . .Do . . 3 1 1 8 Total 34 29 13 1 79 TABLE 116. Cross 55. Rough E (tri) X rough E (tri). RrSS X RrSS = 3 RSS + rrSS (3 E : 1 Sm). No. 9 Rough E. cf Rough E. A B C D E Sm 1 R221 52-12 R140 52-3 ? 3 2 *R201 Sm 52-7 Do ?, Total 4 3 ^ee note, cross 57. TABLE 117. Cross 56. Rough E (tri) X smooth (4-toe). RRSS X rres = RrSs (all C) 1. RrSS X rrss = RrSs + rrSs (1 C : 1 Sm) 2-5. No. 9 Smooth (or rough E). c? Rough E (or smooth). A B C D E Sm 1 399 Sm 4-toe 4003 E Tri 11 2 1 9 Sm 4-toe R140 E 52-3 1 3 3 B31 Sm la-1 .... Do ?, 4 J R201 Sm 52-7 13 W-Sm 4-toe .... 1 1 5 R221 E 52-12 Do ?, Total (1) 11 Total (2-5) 2 8 *See note, cross 57. 154 INHERITANCE IN GUINEA-PIGS. TABLE 118. Cross 57. Smooth (tri) X smooth (4-toe, etc.); both parents of tricolor smooths were partial roughs (cross 52). rr X rr = rr (all Sm). If, however, RSS, normally rough E, is ever Sm: RrSS X rrss = RrSs + rrSs (1 C : 1 Sm). No. ? Smooth. c? Smooth. A B C D E Sm 1 699 R131 52-4 14 2 R139 52-3 99 4-toe 2 3 R13 52-1 Do 8 4 R164 52-13 Do ? 5 R199 Red 52-7 13 W 4-toe 3 6 R249 52-10 Do 3 7 R263 52-11. . . . Do 2 8 X R201 62-7 Do 1 1 *R201 was called rough E? at birth with the note that there seemed to be a trace of roughness on one hind toe. No roughness was apparent when adult and she was called Sm, but nevertheless was tested by mating with a 4-toe smooth. The result shows that she was, genetically at least, like a rough E. TABLE 119. Cross 68. Rough B, C (Lima) X rough B (Lima). No. 9 Rough B. c? Rough B. A B C D E Sm 1 TV7 Tnnnfl. , L5 Lima . . . 9, ? 1 ?, L97 60-6 L26 58-1 1 8 L140 60-7 Do 1 1 4 /L97 60-6 >L98 60-6 ? 9 ? 5 \L81 Red 59-3 L99 Rough C 61-1 . . . Do 1 1 6 LSI, L97, L99 (above) . Do 1 1 2 4 Total 8 7 ?, 1 7 TABLE 120. Cross 59. Rough A (Lima) X smooth (Lima). No. 9 Smooth (or rough A). d" Rough A (or smooth). A B C D E Sm Remarks. 1 L13 Sm 62-2 . . L9 A 58-1 . . ?, 1 ?, L14Sm 60-1 . Do 7 1 ?, 3 L24 Sm 60-2 .Do 9: 1 3 ft Red-B 4 L25 Sm 58-1 .Do 4 1 3 2 Red-A, 2 5 L37 Sm 62-3 . . . . Do 1 Red-Sm 6 L57 Sm 59-3 .Do 1 ? 3 Red-A 7 L24, L57 Sm (above) Do 1 ?, 1 Red-B 8 L22 A 60-2 LI Sm Lima 4 1 6 2 Sep(p)-Sm ft L62 A 59-3 Do 1 ? 10 L100 Sep(p)-A 61-1 L82 Sep(p)-Sm 59-8 ?, 2 Sep(p)-A Total .... 94 8 3 ?3" TABLES. 155 TABLE 121. Cross 60. Rough B (Lima) X smooth (Lima) . No. 9 Smooth. cf Rough B. A B C D E Sm Remarks. 1 L4 Lima . L5 Lima . 3 ?! ? Red-A. 2 Red-B, Red-Sm 2 L6 Sep(p) Lima . Do ?, 1 3 L34 62-2 . L26 68-1 ?, ? 4 L37 62-3 Do 1 5 L43 62-1 . . Do 1 3 fl L34, L43 (above) Do ?, 8 7 L41, L43 62-1 . . Do ?, 3 8 L132 60-3 . . Do 1 1 9 L75 60-4.. Do 1 1 10 f L14 60-1 . . | L24 60-2 [L131 60-3 1 8 Red-Sm [L25 58-1 . . Total 8 9 30 TABLE 122. Cross 61. Rough C (Lima) X smooth (Lima). No. 9 Rough C. d" Smooth. A B C D E Sm Remarks. 1 L56 59-3 LI Lima .... ?! 1 1 ? Sep(p)-A TABLE 123 Cross 62. Smooth (Lima) X smooth (Lima). Offspring all smooth. No. 9 Smooth. c^Smooth. B Red Sep(p) Red(p) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 L3 B Lima . . L2 Red Lima . . LI B Lima . . Do 4 3 4 1 2 3 3 3 1 8 1 7 4 4 6 4 2 1 1 1 1 3 1 1 2 2 4 L8 Red Lima . . Do L2, L8 Red Lima . . Do L12 Red 62-2... Do Lll Red 62-2. . . L40 Red 62-1 . . . L35 Red 62-3 . . . Do L17 Red 62-3 . . . Do L8 Red Lima . . Do L8, L17 Red (above) L64 Red 62-13.. L6 Sep(p) Lima . . L19 Sep(p) 62-3 . . . Do L59 Red 62-3 . LI B Lima . . Do 1 2 L6 Sep(p) Lima . . L19 Sep(p) 62-3. . L18Sep(p) 62-3... . . . . Do Do L82 Sep(p) 59-8... L59 (Red) 62-3... Do 2 3 L58Red(p) 62-3... L121 Red(p) 62-11. . Do L149Red(p) 62-11. . Do L120 Red(p) 62-11. . 156 INHERITANCE IN GUINEA-PIGS. TABLE 124. Cross 63. Rough A, B (Lima) X smooth (4-toe, etc.). No. 9 Rough A (or smooth) . c? Smooth (or rough A). A B C D E Sm 1 L107 A 59-9 .... 13W-Sm 4-toe 9, 1 2 L108 A 59-7 Do 1 3 L110B 59-7 Do 1 T 4 D180 W-Sm 18c-14. . L98 B 60-6 ?, 1 5 D236 W-Sm 17d-7. . . Do ?, 1 Total . . . 4 4 4 TABLE 125. Cross 64. Rough C (low grade due to C. rufescens) X smooth (4-toe, etc.). RrSs X rrss = Rrse + RrSs + 2rr(lA:lC:2 Sm). No. 9 Rough C. cf Smooth. A B C D E Sm Remarks. 1 A606 Ag \ . . . . 166 .4-toe... 2 2B-C 2 A1687 64-1 . 99 4-toe . . . 1 B-Sm 3 A1688 Ag 65-1 AA83 fa 3 4 BrAgTb-A, 2 AgTb-A, 4 Total . . s f, 5 AgSm TABLE 126. Cross 65. Smooth (some C. rufescens blood) X rough A. rrSs X Rrss = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). rrss X Rrss = Rrss X rrss (1 A : 1 Sm). No. 9 Smooth. cf Rough A. A B C D E Sm 1 A702 AgTb fa 2597 Ag stock 1 1 2 A605 i .. .Do. 1 1 3 A642 j Do ?, 4 A842 1 Do 2 3 fi A913AgTb fa Do 2 6 699 AgTb T i-iif. Do 7 }?, 7 B238 la-5 .... R88 52-1 9 4 8 fB240 la-5.... \ Do . . 2 5 \K61, K62 78-6.... / TABLES. 157 TABLE 127. Cross 66. Rough A X smooth; both parents with a little C. rufescens blood. Rrss X rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Rough A. cfSmooth. A B C D E 3m Remarks. 1 A1690Ag 65-5.. AA83 B-Sm ^j . . . 4 1 4 3 AgTb-A, Red-A, Ag ?, Do M91 Ag-Sm 8-4.. ? 4 Tb-B, AgLb-Sm, Ag Tb-Sm, 2 Red-Sin 2 AgLb-A, 2 AgLb-Sm, a A1691 Ag 65-5.. AA83 B-Sm ^ 8 4 2 AgTb-Sm 4 AgLb-A, 4 B-A, 2 Ag Total 14 1 1? Lb-Sm, 2 B-Sm TABLE 128. Cross 67. Rough A (4-toe, tri) X smooth (pure lea). Rrss X rrSS = RrSs + rrSs (1 C : 1 Sm). Young all light-bellied agouti. No. 9 Rough A. d* Smooth. A B C D E Sm 1 2 3 4 /R215 49-1 .... \R252 49-2.... R236 50-3 J724 SAg(R) lea .... 3 Do 2 1 1 2 4 R205 46-4 . . . Do R213 49-1 . . Do Total 6 TABLE 129. Cross 68. Rough A X smooth (pure C. cutleri). Rrss X rrSS = RrSs + rrSs (1 C : 1 Sm). Young all light-bellied agouti. No. 9 Rough A. cfSmooth. A B C D E Sm 1 3986 4-toe C128 Ag .... 1 ?, 2 3988 4-toe... Do 2 1 3 3986,3988 4-toe... Do 2 3 4 A1691 Ag 65-5 . . . Do 1 2 3 5 AA567Ag 66-3... Do 1 3 6 /AA568 66-3... \R80 54-15.. J....DO ... 3 Total 3 6 15 158 INHERITANCE IN GUINEA-PIGS. TABLE 130. Cross 69. Rough C (tricolor) X smooth (pure C. cutleri). RrSs X rrSS = RrSs + RrSS +2rr(lC:lE:2 Sm). Young all light-bellied agouti. No. ? Black rough C. D. cf Smooth. A B C D E Sm 1 3245 Tri C128Ag ? 2 R54 52-1 . . . Do 1 1 1 3 R152 54-4 . . . Do 1 ? 4 3245, R110 Do ?, 1 5 R110 51-1... Do ?, 1 6 R170 47-3 . . . Do 1 a 7 Rll 52-1... Do 1 8 RIGID 52-5... Do 3 9 R154 D 54-4 . . Do . . . . a Total ... . 1 1 q T> TABLE 131. Cross 70. Rough A (guinea-pig) X rough C, D (J, J cutleri). Rrss X RrSs = Rss + RSs + 2rr (3 A : 3 C : 2 Sm). All | cutleri Rough C, D, except K58, J blood. R116 and R137 may be RRsa. No. 9RoughA(orC,D). d" Rough D(orA). A B C D E Sm Remarks. 1 R116B-A 46-4.. K54 Ag-D 68-1 . . a 1 Ag-A, B-A, Ag-C 2 R117B-A 46-4.. Do 1 1 B-D, B-Sm 3 R137 B-A 47-1 . . Do ?: 2B-C 4 AA608 B-A 66-3 . . Do 3 1 Ag-A, 2 B-A, Ag-E 5 K12 Ag-D 68-3 . . R31 B-A 45-3.. a 1 2 B-C, B-Sm 6 K14 Ag-D 68-3. . Do a 2 Ag-B 7 Do 3609 B-A 4-toe . 1 i 1 B-A, Ag-B, Ag-E 8 Do R76 B-A 45-4 . 1 1 B-A, B-D 9 K12Ag-D 68-3 . Do 1 ? 1 Ag-A, 2 Ag-C, B-Sm 10 K58B-C 70-5. Do i 1 B-C, B-Sm Total . . . 8 3 8 9 a 4 TABLE 132. Cross 71. Rough A (guinea-pig) X smooth (5, J cutleri). Rrss X rrSs = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). All i cutleri except K79, \ cutleri. No. 9 Smooth. cf Rough A. A B C D E Sm Remarks. 1 K7Ag 77-1 . R31 B 45-3. 3 ? 1 7 Ag-A, 2 B-A, 2 B-B, B-C, 6 2 K15Ag 68-3.. Do 3 ? q Ag-Sm, B-Sm Ag-A, 2 B-A, Ag-C, B-C, 3 3 K55 Ag 68-1 . Do ? 1 Ag-Sm, 6 B-Sm 2 Ag-C, B-Sm 4 K7, K55 (above) Do ? 3 2 Ag-A, 2 Ag-Sm, B-Sm 5 K68 Ag 77-1 . Do 1 1 ? 1 B-A, B-B, Ag-D, 2 B-C 6 K116 Ag 68-6 Do 1 B-Sm 7 K81 Ag 69-1 . 3609 B 4-toe . 1 1 Ag-C, Ag-Sm 8 K79 B 78-1 . 3922 B 4-toe . 1 1 B-A, B-C Total 10 3 q 1 ?? \ TABLES. 159 TABLE 133. Cross 72. Smooth (guinea-pig) X rough C, D ($, J cutleri). rres X RrSs = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). K71a, K92 may be RR. No. 9 Rough C, D(orSm). cf Smooth (or rough C, D). A B C D E Sm Remarks. 1 K12 Ag-D 68-3 . . . OOCr(Br)Sm Dil 1 2 Ag-A, 2 B-D ?, K14 Ag-D 68-3 . . . Do 3 6 2 Ag-A, B-A, 3 3 fK12 Ag-D 68-3 } Do . . 1 3 Ag-Sm, 3 B- Sm fB-C, 2 Ag-Sm, 4 \A71a Ag-C 70-1 . . . K92 Ag-C 70-9 . . . / Do 1 7, \ B-Sm Ag-A, 2 Ag-C 5 K157Ag-C 72-12.. 13Cr(Br)SmDil... . 1 1 Ag A, Sep Sm 6 K147 AgTb-D 72-13 . . Do 3 2 B-Sm,Sep-Sm 7 K142 B-D 72-1 . . . OOCr(Br)SmDil... . 1 1 1 Red-A, W-D, 8 D40Cr(Br)Sm 36-1... K60 Ag-C 71-2 . . 1 1 B-Sm B-A, Ag-Sm 9 R173 B-Sm 49-1 . . . Do 1 1 Ag A, Ag Sm in 20 B-Sm 4-toe . . K54 Ag-D 68-1 . . ?, 1 2 Ag-A, B-A, Ag- 11 AA533Ag-Sm 66-2... K56B-C 70-5.. 1 C, 2 B-Sm Ag-Sm i?i AA586 Ag-Sm 106-10 . Do 1 1 Ag-C, B-D 13 fM382 AgTb-Sm 16-9. . . Do 1 1 2 1 f B-A, 3 AgTb- 14 \B239 AgTb-Sm la-5 . . . B239 AgTb-Sm la-5 . . . Do 1 \ C+D,B-Sm B-Sm Total 1? 6 6 ?1 TABLE 134. Cross 73. Smooth (4-toe) X rough A, B (J cutleri). rrss X Rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Rough A, B. c? Smooth. A B C D E Sm Remarks. 1 K95 B-B 71-1 99 B 4-toe 1 B-Sm 2 fKlOlB-A 70-8 J13W 4-toe... ?, 1 2 B-A, B-Sm \K106 B-A 71-2 Total J ? ? TABLE 135. Cross 74- Smooth (, \ cutleri) X rough A rrSs X Rrss = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). rrss X Rrss = Rrss -j- rrss (1 A : 1 Sm). cutleri). No. 9 Smooth. cf Rough A. A B C D E Sm Remarks. 1 K68 Ag 77-1 K59 B 71-2 . . 1 B-C 2 K42 B 78-2 Do 1 B-Sm Total 1 1 160 INHERITANCE IN GUINEA-PIGS. TABLE 136. Cross 75. Rough A, B (\ cutleri) X rough A ( cutleri). Rrss X Rrss = 3Rss + rrss(3A:l Sm). No. 9 Rough A, B. cT Rough A. A B C D E Sm Remarks. 1 K50 Ag-B 70-6 K59B 71-2 1 B-A ?, K53 B 71-1 Do ? 2B-A Total s TABLE 137. Cross 76. Rough C (f cutleri) X rough C (i cutleri). RSs X RSs = Has + 2 RSs + RSS (1 A : 2 C : 1 E : ? Sm ?). No. 9 Rough C. c? Rough C. A B C D E Sm Remarks. 1 K114B 74-1 K93 Ag 70-9 . . 1 3 1 fAg-C, Ag-D, 2 B-D, I Ag-E TABLE 138. Cross 77. Black (BB) X agouti (pure cutleri). Parents and offspring all smooth. No. 9 Black. c? Agouti. AgLb Black W 1 2 B 9 9 BB C128Ag pureC 6 TABLE 139. Cross 78. Black (BW) X agouti^, \, J cutleri). Expectation : 1 Ag : 1 black : some white. Parents and offspring all smooth. No. 9 Black (or agouti) . % cf Agouti (or black). AgLb Black W 1 6B 9 9 BW C67 Ag (i) - . . . 12 12 6 ?: 2B 9 9 BW K4Ag(J) 78-1 6 9 3 4 3B 9 9 BW K20 Ag (}) 78-1 K24 Ag (i) 78-1 .... 39 B BW 9 3 5 fi K22 Ag (}) 78-1 Do 1 1 6 K29 Ag (i) 78-1 Do 3 3 7 K103Ag(i) 78-4... Do 2 8 K66Ag(i) 78-2. . Do 4 9 K109Ag(i) 78-6. . Do 1 1 in 3B 9 9 BW.. . K104Ag(J) 78-4 5 3 Total 35 41 10 PART III FURTHER STUDIES OF PIEBALD RATS AND SELECTION, WITH OBSERVATIONS ON GAMETIC COUPLING BY W. E. CASTLE THE PROGENY OF HOODED RATS TWICE CROSSED WITH WILD RATS. In 1914 Castle and Phillips published a report on breeding experi- ments with hooded rats, in which it was shown that the hooded color pattern itself a Mendelian recessive character in crosses with the entirely colored (or "self") coat of wild rats is subject to quantitative variation, and that different quantitative conditions of the hooded pattern are heritable. (Compare fig. 36, plate 7.) It was also shown that by repeated selection of the more extreme variations in the hooded pattern (either plus or minus) it is possible gradually to modify the racial mean, mode, and range as regards these fluctuations, without eliminat- ing further fluctuation or greatly reducing its amount. We concluded that the unit character, hooded color pattern, is a quantitatively vary- ing one, but were at that time unable to decide whether the observed variability was due simply and exclusively to variation in a single Mendelian unit factor or partly to independent and subsidiary modify- ing Mendelian factors! Since publication of the above I have been engaged in further experi- ments designed to show which of the alternative explanations is the correct one, and these are now sufficiently advanced to indicate definite conclusions. Previous experiments had shown that when a race of hooded rats, whose character has been modified by selection (either plus or minus), is crossed with wild rats, the extracted hooded animals obtained in F 2 as recessives show regression toward the mean condition of the recessive race before selection began. This result suggested that the regression observed might be due to removal by the cross of modifying factors, which selection had accumulated in the hooded race. If this view was correct, it was thought that further crossing of the extracted hooded animals with the same wild race should result in further regression, and that if this further regression was not observed a different explanation must be sought for the regression already noted. The entire experiment has accordingly been repeated from the beginning, with the same result as regards regression in the first F 2 generation, but with no regression of the same sort in a second F 2 con- taining twice-extracted hooded animals. So far from observing further regression as a result of the second cross with wild rats, we have unmis- takable evidence that the movement of the mean, mode, and range of the hooded character has been in the reverse direction. So the hypothe- sis of modifying factors to account for the regression and for the pro- gressive changes observed under selection becomes untenable. In repeating the experiment of crossing hooded rats of our selected races with wild rats, great care has been taken to employ as parents individuals of the greatest racial purity and to inbreed the offspring 163 164 INHERITANCE IN RATS. brother with sister, thus precluding the possibility of introducing modifying factors from other sources. In making the second set of crosses, the extracted individual has, wherever possible, been crossed with its own wild grandparent. In the few cases hi which this was impossible, wild animals of the same stock have been used. This stock consisted of a colony of wild rats which invaded the basement of the Bussey Institution apparently from a nearby stable. Owing to faulty construction of the building they were able to breed in spots inaccessible to us, and it took many months of continuous and persistent trapping to secure their extermination. During this period we trapped a hun- dred or more of them, all typical Norway rats, colored all over, without even the white spot occasionally seen on the chest of wild rats. Two generations of rats from this wild stock have been reared in the labora- tory, and all have this same self-colored condition. The hooded animals used in the experiments to be reported on in this connection consisted of 4 individuals of the plus-selected series, a male and 3 females, as follows: TABLE 140. Individual. Grade. 1 Generation. 95513 +4i 10 d"6348 +4 10 96600.... +4i 12 96955.... +4 12 ^ee figure 35, plate 7, for significance of the grades. Each of these animals was mated with a single wild mate, and their children were weaned directly into breeding cages containing a male and two or three females (brother and sisters). In the case of two matings, F! males of the same parentage were at the time lacking and males from a different cross were used. The results of such matings are tabulated by themselves and serve a useful purpose as controls. The F! animals all closely resembled their wild parents, but many of them had a white spot on the chest. They ranged from grade +5 to +6 (self). The F 2 animals are classified in table 141, where it appears that 73 of them were hooded and 219 non-hooded (i. e., like F x ), an exact 1 : 3 ratio. More than half of this F 2 generation consists of the grand- children of 95513, produced by breeding her children brother with sister, those children all having been sired by the same wild rat. Her grandchildren include 41 hooded and 107 non-hooded young. The hooded young range in grade from + 1 to -f-4, their mean grade being +3.05, a considerable regression from the grade of the grandmother, which was 4.25. Hooded rats of the same grade and generation as the grandmother, when bred with each other, produced young of mean grade +3.84. HOODED CROSSED WITH WILD. 165 (See table 10, Castle and Phillips.) The mean of the extracted hooded grandchildren in this case (being 3.05) shows a regression of 0.79 from that expected for the uncrossed hooded race. From the extracted hooded grandchildren of 9 5513, produced as just described by a cross with a wild male, 7 individuals, 2 males and 5 females, were selected for a second cross with the wild race. They ranged in grade from +2 to +3|. (See table 142.) They produced several litters of young of the same character as the first F! young, all being similar to wild rats in appearance, except for the frequent occurrence of a white spot on the belly. These second F! young were at weaning time mated, brother with sister, in breeding-pens, precisely as had been done with the first FI'S. They produced 394 second F 2 young, of which 98 were hooded and 296 non-hooded, a perfect 1 : 3 ratio. The hooded young varied in grade from +2 to +4 ; as shown in table 142, the data there being given for each family separately as well as for all combined in the totals. One family was very like another as regards the character of the hooded young, except that the higher-grade grandparents had grandchildren of slightly higher grade. Thus the average of all the 98 hooded young was +3.47, but the average of those descended from the 3 grandparents of lowest grade was less than this, while the average of those descended from the 3 grandparents of highest grade was greater. This is just what had been observed throughout the entire selection experiments. (See Castle and Phillips.) If we weight each of the grandparents in table 142 in proportion to the number of its hooded grandchildren, then the mean grade of all the grandparents is +2.95. Since the mean grade of all the 41 first F 2 hooded grandchildren, from which these 7 were chosen, was +3.05, it will be seen that these 7 are, so far as grade is concerned, fair repre- sentatives of the 41, being in fact of slightly lower mean grade. It is therefore all the more striking that their grandchildren, the second F 2 hooded young (table 142) , are of higher grade. They regress in an oppo- site direction to that taken by the first F 2 hooded young. Thus the original hooded ancestor ( 9 5513) was of grade 4.25. The grade of hooded young expected from such animals is 3.84. What she produced in F 2 , following a cross with the wild male, was young of mean grade 3.05. Seven of these of mean grade 2.95 produced a second F 2 contain- ing hooded young of mean grade 3.47. This is a reversed regression of 0.52 on the grade of their actual hooded grandparents, or of 0.42 on the group from which their grandparents were chosen. Their mean lies about midway 1 between that which would have been expected from the original hooded female (5513) had no crossing with wild rats occurred and that which was observed in the first F 2 . 1 In The Scientific Monthly (Jan. 1916) I have stated that a second cross showed "a return to about what the selected race would have been had no crossing at all occurred." This is obviously inaccurate and should be corrected. It rests on a comparison with the combined average of both the older and the more recent experiments. 166 INHERITANCE IN RATS. Obviously these facts do not harmonize with the assumption that the regression observed in the first F 2 was due to loss of modifying fac- tors accumulated during the ten preceding generations of selection; for no further loss occurs in the second F 2 . On the other hand, a partial recovery is made of what was lost in the first F 2 . This suggests the idea that that loss may have been due to physiological causes non- genetic in character, such as produce increased size in racial crosses; for among guinea-pigs (as among certain plants) it has been found that Fj. has an increased size due to vigor produced by crossing and not due to heredity at all. This increased size persists partially in F 2 , but for the most part is not in evidence beyond FV I would not suggest that the present case is parallel with this, but it seems quite possible that similar non-genetic agencies are concerned in the striking regression of the first F 2 and the subsequent reversed regression in the second F 2 . Whatever its correct explanation may be, the fact of the reversed regression in a second F 2 is very clear, as other cases than those already discussed will show. A hooded rat of grade +4 and generation 10, c?6348, had by a wild female several young of the character already described for the young of 9 5513. These, mated brother with sister, produced a first F 2 (table 141) of 90 rats, 22 of which were hooded, 68 being non-hooded, again a good 1 : 3 ratio. The hooded young ranged from +2 to +4 in grade, their mean being 3.28. Of the 22 hooded individuals, 1 male and 7 females were mated with wild rats to obtain a second F 1; and the second FI animals were then mated brother with sister to obtain the desired second F 2 . The character of this is shown family by family in table 143. It contained 497 individuals, of which 121 were hooded and 376 non-hooded, a ratio of 1 : 3.1. The weighted mean of the 8 selected grandparents is 2.93, which is 0.35 below the mean of the 22 first F 2 hooded animals which they represent. The mean of the second F 2 hooded young is 3.22, which indicates a reversed regression of 0.29 on the grade of the grandparents, but shows no significant difference from the mean of the grandparental group (3.28). All except one of the 8 families classified in table 143 show unmis- takably the reversed regression. This exceptional family consists of the grandchildren of 9 9747. They have a mean grade of 2.90, sub- stantially the same as that of the entire group of grandparents but con- siderably lower than that of their own hooded grandmother. Appa- rently she did not come up genetically to her phenotypic grade. This the other grandparents of the group did. For those of lowest grade (2, 2f ) produced lower-grade hooded grandchildren than did the grand- parents of highest grade (3|, 4), as was found to be the case also, in table 142. We may next trace the inheritance of the hooded character through a third but smaller family produced by two successive crosses with wild HOODED CROSSED WITH WILD. 167 rats, the hooded character in this case being derived from 96955, grade +4, generation 12. The character of her first F 2 descendants is shown in table 141. They consist of 5 hooded and 27 non-hooded individuals. The mean grade of the hooded young is 3.51, but the number of these young is too small to make this mean of much signifi- cance. One of the hooded young (cT9660,+3f ) was mated with a wild female to secure a second F! generation and from this in due course was produced the second F 2 generation (table 144) . It consisted of 21 hooded and 44 non-hooded young. The hooded young showed the usual range (2 to 4). Their mean grade was 3.50, substantially identical with that of the first F 2 animals, but 0.25 below that of the actual hooded grand- parent. This family history is less satisfactory than the two already discussed because of the smaller numbers which it includes. It con- tains nothing contradictory to the interpretation already given, though reversed regression is not in this case in evidence. In two cases FI females could not be mated with brothers and so mates were taken from other families. Thus " mixed FI matings" were made between children of 5513 and 6600 and children of 5513 and 6955. (See table 141.) The former mating produced 3 hooded and 12 non-hooded "first" F 2 young; the latter produced 2 hooded and 5 non-hooded "first" F 2 young. The grade of the hooded young pro- duced by these mixed matings was not different from that of brother- sister matings, so far as the small numbers permit one to judge. One of these mixed matings was carried into a second F 2 generation. The first F 2 hooded cT9711, +3|, was mated with a wild female, and the young were bred, brother with sister, producing 16 hooded and 33 non- hooded young. (See table 144.) The mean grade of the 16 hooded young was 3.28, nearly the same as that of the first F 2 hooded grand- parent. No additional regression through loss of modifiers (or other agency) is here in evidence. The result is the same as that observed in families wholly unmixed. The attention of my pure-line critics, who think that in our mass-selection experiments insufficient attention has been given to individual pedigrees, is particularly directed to the foregoing case. Having now discussed each family history separately, we may com- bine all the second F 2 families in one table, in order to get a clearer impression of the results as a whole. (See table 145.) The second F 2 generation thus combined includes 256 hooded and 749 non-hooded individuals, a ratio of 1 : 2.9, an unmistakable mono-hybrid Mendelian ratio. The mean grade of the hooded individuals is 3.34. The weighted mean grade of their hooded grandparents was 3.02, which indicates a reversed regression of 0.32 for the entire second F 2 group of hooded animals. Classified according to the grade of the (first F 2 ) grandparent, they show a correlation between grade of grandparent and grade of grand- 168 INHERITANCE IN RATS. child. The lower-grade grandparent has lower-grade hooded grand- children, and the higher-grade grandparent has higher-grade hooded grandchildren. This shows that the variation in grade is (in part at least) genotypic. As the experiment yields no evidence that the varia- tion in the hooded character is due to independent modifying factors, there remains no alternative to the conclusion that the single genetic Mendelian factor concerned fluctuates in genetic value. Fluctuation accordingly is not exclusively phenotypic, as DeVries and Johannsen have thought, but may be genetic also. Hence racial changes may be effected through selection by the isolation of genetic fluctuations, as well as by the isolation of mutations. Moreover, genetic fluctuation makes possible progressive change in a particular direction, repeated selection attaining results which it would be quite hopeless to seek by any other means. A SECOND REPORT ON MASS SELECTION OF THE HOODED PATTERN OF RATS. The experiments in selection for the modification of the hooded pat- tern of rats, when reported on by Castle and Phillips in 1914, had been carried through 13 generations. Since then the experiments with the same selected races have been carried through 3 or 4 additional genera- tions, the results of which will now be described. Additional records have also been obtained for certain of the generations reported on by Castle and Phillips, which may now be combined with those previously pub- lished. Thus, revised data, based on larger totals, may be given for generations 12 and 13 of the plus-selection series and for generation 13 of the minus-selection series. These do not materially change the results previously obtained, but add to their trustworthiness. The additional generations of selection show a continued progressive movement of the racial character in the direction of the selection and indicate the exist- ence of no natural limit to the progress which selection can make in changing the hooded character. For details concerning the earlier history of the experiments and the methods of grading the animals the reader is referred to the publica- tion of Castle and Phillips. The grading scale (exclusive of the newer and more extreme grades) is reproduced in figure 35, plate 7. Atten- tion may be called to the fact that the entire selection series, both plus and minus, consist of animals descended from an original stock of less than a dozen individuals. These descendants number more than 33,000. In their ancestry, since the beginning of the selection experiment, not a single cross out of the race has occurred. At the same time no effort has been made to avoid inbreeding. Brother and sister and cousin matings are frequent in our records. Under these circumstances it is inevitable that the selected races should have become much "inbred." MASS SELECTION. 169 Our critics with a leaning toward the "pure-line" idea have insisted that nothing but brother-sister matings should have been employed in our experiments. We have several times endeavored to carry forward certain high-grade families on this basis, but have been unable to secure large enough numbers of offspring to make this possible; but we have in several cases produced families of considerable size, descended exclu- sively from a single pair of ancestors notably in the case of our pure "mutant" race and in a race descended from one hooded and one wild rat, which race was continued through 8 filial generations. (See p. 21, Castle and Phillips.) It would have been impossible, in these and other races, to make as rapid progress as we secured through selection in our two principal races, for when only brother-sister matings are permitted, it often happens that a mate of proper grade can not be secured for an individual among its own brothers and sisters, though such a mate may be found among its cousins or more remote relatives. It being our first object to test the effectiveness of selection, we have made selec- tion of any individual within the group (series or family with which we were dealing) regardless of relationship, making the selection as rigid as the maintenance of a stock of considerable size would permit. More than once we have crossed the danger-line in advancing the standard of selection to such an extent that only small numbers of parents came up to it; more than once we have had to relax our standard temporarily in order to keep the race alive. That the long-continued inbreeding of our selected races has affected their vigor and fecundity is unquestionable. It is shown by the fact that the plus and minus races, which had a common origin many generations ago and have ever since been bred in the same room and under identical conditions, if crossed with each other, produce offspring of much greater vigor and fecundity than either parent strain. In this our observations on the effects of inbreeding are entirely in harmony with those of Darwin, Bos, Weismann, and of breeders of farm animals quite generally. Miss King is credited with the view that inbreeding of rats may increase their size, vigor, and fecundity, but this is cer- tainly contrary to common experience with these and other animals. It is probably true that under inbreeding it is possible, in exceptional cases, to isolate a strain relatively immune to ill effects from inbreeding (like Darwin's "Hero" morning-glory) or so inherently vigorous that it succeeds in spite of inbreeding. But it is very doubtful whether inbreeding of itself affects vigor other than disadvantageously. It is a sufficient test to cross-breed an inbred strain, in order to ascertain whether the inbreeding has increased or impaired its vigor. 170 INHERITANCE IN RATS. PLUS AND MINUS-SELECTION SERIES. The plus-selection experiment, when described by Castle and Phil- lips, had been carried through 13 generations, but the last 2 genera- tions were incomplete. The number of offspring included in genera- tion 12 (table 146) has now been raised from 590 to 682 and the number of offspring included in generation 13 (table 147) has been raised from 194 to 529. The mean grade of the parents for generation 12 has advanced from 4.09 to 4.10; that of the offspring has fallen from 3.94 to 3.93. Neither of these changes is of significant size. The correla- tion is now found to be 0.168 instead of 0.161. In generation 13 (table 147) the changes are greater, as might be expected from the greater change in the number of observations. The mean of the parents is now 4.13 (formerly 4.22) ; that of the offspring is 3.94 (instead of 3.88). The correlation is 0.117, as compared with 0.132, the value previously obtained. Generation 14 (table 148) includes 1,359 offspring of mean grade 4.01. They are descended mostly from parents of grade +4 or higher, mean 4.14. Generation 15 (table 149) includes 3,690 individuals, more than have been produced hi any other generation of the series. The mean grade of the parents was in this generation advanced about a quarter grade to 4.38; that of the offspring advanced a little, to 4.07. Generation 16 (table 150) was also large, including 1,690 offspring. The grade of the parents was again advanced a little to 4.45; that of the offspring followed a similar amount, to 4.13. In the three generations (14 to 16) which have been added since the last report, the grade of the selected parents has been advanced by 0.32, from 4.13 to 4.45; that of the offspring has advanced 0.19, from 3.94 to 4.13 (the mean grade of the parents three generations earlier). The upper limit of variation of the offspring has meanwhile advanced from 5.25 to 5.87, the highest grade being found in a rat black all over except for a few white hairs on the chest. This rat has produced a few offspring of almost as high grade, though the most of his young are of much lower grade. In the minus-selection series, generation 13, in our previous report, contained 571 offspring. This number has now been raised to 1,006 (table 151), the mean grade of both parents and offspring being prac- tically unchanged by the additional young recorded. The parents are of mean grade 2.49, the offspring of mean grade 2.40. In the next generation (14) the offspring number 717, their mean grade being 2.48, that of the selected parents being 2.64. (See table 152.) Generation 15 includes 1,438 young of mean grade 2.54. The mean grade of the parents is 2.65. (See table 1530 N * MASS SELECTION. 171 Generation 16 is the largest in the minus-selection series. It includes 1,980 young of mean grade 2.63. The grade of the parents is 2.79. (See table 154.) Generation 17 (table 155) includes 868 young of mean grade 2.70. The grade of their parents is 2.86. Four generations of selection have thus been added to the minus series as it stood at the last report. The mean grade of the parents has been advanced from 2.49 to 2.86; that of the offspring from 2.40 to 2.70, the former is an advance of 0.37, the latter of 0.30. In the plus series the corresponding changes for one less generation of selection (three), were 0.32 and 0.19, respectively. In both series a change in the mean of the offspring attends that in the parents, coin- ciding with it in character but not quite equaling it in amount. The lagging behind of the offspring, as compared with their selected parents, gives a good illustration of regression, the phenomenon made familiar by Galton's researches, but explained away by Johannsen as due to a sorting-out action of selection on mixed races. The extent to which in these experiments the offspring lag behind their parents or "regress on their parents" is indicated in each table in the column headed " regression." Tables 146 and 150 illustrate particularly well how the offspring regress toward the general average of the race for the time being. The offspring of parents substantially the same grade as the general average of the race show no regression; the offspring of parents below this average show regression upward (indicated hi the tables by the minus sign); the offspring of parents above the racial average show regression downward, the amount of the regression increas- ing with the aberrant character of the parents. If one examines either selection series as a whole (compare Castle and Phillips), he will notice that the point (toward which regression occurs) changes with the progress of the selection. At the beginning of the plus-selection series regression was toward a grade of about 4-1.75 (see table 1, Castle and Phillips); after about 15 generations of plus selection it has advanced to +4.00. (See tables 148 to 150.) At the beginning of the minus-selection series, regression occurred toward a grade of to 1 (Castle and Phillips, tables 16 and 17); in generation 17 (table 155) regression is apparently toward grade 2.62. These grades toward which regression occurs represent points of racial equilibrium or stability at which the race would tend to remain in the absence of further selection, but these points of equilibrium are capable of being moved either up or down the scale of grades at the will of the breeder, provided he has patience and persistency and will select repeatedly. Regression indicates that there is not complete agreement between the somatic and the genetic character of the parents selected. But the steady movement (in the direction of the selection) of the point of 172 INHERITANCE IN RATS. equilibrium toward which regression occurs serves to show that geno- typic as well as phenotypic fluctuations occur in the material on which selection is brought to bear. DeVries and Johannsen have damned the word fluctuation by ascribing to it purely phenotypic sig- nificance. Is it not worth while to rescue the term from its present odious position, since it is clear that variation having a genetic basis may in every way resemble somatic fluctuation, except in its behavior under selection? Fluctuation may conceivably be either somatic or genetic or both. No one, in advance of actual experiment, can tell what its nature is in a particular case. In the case under discussion the fluctuation is obviously partly somatic and partly genetic. The somatic fluctuation occasions regression, the genetic fluctuation per- mits a change (under selection) of the point toward which regression occurs that is, in the general average of the race. Tables 156 and 157 show (generation by generation) the progress made by selection in modifying the racial character. It will be observed that as the mean advances in the direction of the selection both the upper and the lower limits of variation move in the same direction. The amount of the variation as measured by the standard deviation is less in the last half of the experiment than in the first half. It is also steadier, owing in part doubtless to the fact that the numbers are larger, and in part to a more stable genetic character of the selected races. But the genetic variability is plainly still large enough to per- mit further racial modification and there is no indication that it will cease until the hooded character has been completely selected out of existence, producing at one extreme of the series all-black rats, and at the other end of the series black-eyed white rats. FURTHER OBSERVATIONS ON THE "MUTANT" SERIES. Castle and Phillips described, under the name "mutants," 2 rats of the plus-selection series of very high grade. They proved to be heterozygotes between the average condition of the plus-selected race at that tune, about +3.75, and a new condition, not previously known in our hooded races, but resembling that seen hi " Irish " rats, which are black all over except for a white spot on the belly and would be classed on our grading scale as about +5. In later generations we secured animals homozygous for the darker condition just described (that of Irish rats). The homozygous "mutant" race proved to be very stable hi color-pattern, varying only from 5| to 5f, with a majority of ani- mals graded 5|. Attempts to alter the modal condition of the race by selection have thus far proved futile because of our inability to increase the race sufficiently to afford a basis for selection. Its inbred- ness and its feebleness are perhaps causally related. The suggestion was made that the change from our plus-selected race, which had occurred hi the mutant stock, might be due to some supplementary modifying factor, not to a change in the hooded factor itself. If so, a cross with a race lacking the hooded factor or its "modi- fiers" might serve to demonstrate then* distinctness by separating one from the other. A wild race seemed best suited for a test of this hypothesis, since it would be free from suspicion on the possible ground of harboring either the hooded pattern or its supposed modifier, which had converted the hooded pattern into the mutant. It was to be expected, if the hypothesis were correct, that the mutant character was hooded plus modifier; that then a cross with wild should produce in F 2 hooded young (lacking the modifier) as well as mutants and selfs. But if the mutant race had arisen through a change in the hooded factor itself, then the cross should produce only mutants and selfs, without hooded young in F 2 . Crosses have now been made on a sufficient scale to show beyond question the correctness of the latter alternative, which is entirely in harmony also with the results described in the preceding parts of this paper. Six homozygous "mutant" females of grade +5^ were mated with wild males of the same race described in Part I. They produced 46 young, all gray like wild rats and of grades as follows: Grade 5 5f 5| 6 No 1 15 7 23 Exactly half of the 46 F! rats bore no white spot, i. e., were of grade +6. Seven more bore only a few white hairs (grade 5|) . The remain- der were very similar to the mutant parent in grade. Several matings were made of the F x rats, brother with sister, which produced 212 F 2 young. About a quarter of these were black (non- 173 174 INHERITANCE IN RATS. agouti), the rest being gray (agouti). Both sorts included about equal numbers of individuals with and without white spots. No difference was observed in this respect between the progeny of spotted and of unspotted parents. Table 158 shows the F 2 young grouped family by family according to grade. Three of the four families are descended from a single mutant grandparent; the fourth family is descended from two different mutant grandparents which were bred simultane- ously to the same wild male in the same cage. The 10 F 2 young of this family may have been produced either by full brother and sister, or by half-brother and half-sister; it is uncertain which. All other F 2 young were produced by brother-sister matings. It will be observed that the F 2 young (table 158) which are white- spotted are in no case hooded. Their range of variation does not fall beyond that of the uncrossed mutant race. It is certain, therefore, that the "mutant" condition is not hooded plus an independent Mendel- ian modifier. It is a changed form of white-spotting, alternative to the form of spotting found in the race from which it was derived (the plus- selection series, generation 10). It is, without much doubt, also alter- native to the self condition of wild rats, though fluctuation in grade ob- scures the segregation, which may, very likely, be imperfect. This serves to confirm the general conclusion that throughout the entire series of experiments with the hooded pattern of rats we are dealing with quan- titative variations hi one and the same genetic factor. GAMETIC COUPLING IN YELLOW RATS. Two yellow-coated varieties of the Norway rat (Mus norvegicus) made their appearance as sports or mutations in England a few years since (Castle, 1914) and are now recognized as distinct varieties by fanciers. Both are similar in appearance except for the eye color. In one variety the eye is pink, showing under gross inspection only the color of the blood in the retina. In the other variety the eye is a reddish-black, owing to the combined effect of the red-colored blood and the black-pigmented retina. Since the retinal pigment is much less in this variety than in rats with gray or black coats, the eye is redder. It will be convenient to distinguish the dark-eyed yellow variety as red-eyed, reserving the name black-eyed for gray or black rats. In the coats of both the pink-eyed and the red-eyed varieties of yellow rats black pigment is very feebly developed. It is in fact of a pale cream color. But the true yellow pigment seen on the tips of the hairs of gray rats is retained in full intensity in the yellow varieties. For this reason agouti varieties of yellow rats are much brighter-colored than non-agouti varieties. A non-agouti yellow variety has fur cream- colored throughout its length; the corresponding agouti variety has fur of this same cream color at its base, where the fur of gray rats is black- pigmented, but the hair-tips are of a bright yellow color of exactly the same shade as the hair-tips of gray rats. Hence it is clear that in these yellow varieties of rats a genetic factor for black pigmentation has been affected without any apparent change in the genetic apparatus for producing ordinary yellow pigment. This is quite different from the genesis of yellow coat in most ro- dents for example, in guinea-pigs and rabbits in which black pig- ment is not apparently changed in character but merely in distribution, being "restricted" chiefly to the eye. In the yellow varieties of rats black pigment seems to be affected in the same way as in the pink-eyed variety of guinea-pigs and mice, viz, to be greatly weakened without affecting in the least the development of yellow pigment. The genetic behavior as well as the appearance of the pink-eyed yellow variation in rats is in every way parallel with the behavior of the variations known by the same name hi mice and guinea-pigs. But red-eyed yellow in rats is a genetically distinct variation, as we shall presently see. In no other mammal does there occur a parallel variation, so far as I know. Both red-eyed yellow and pink-eyed yellow were found to be recessive Mendelian variations in crosses with black-eyed rats. From a cross between black-eyed and red-eyed an F 2 generation of 4 609 rats was raised, of which 452 were black-eyed and 157 red-eyed; expected, 457 : 152. From a cross between black-eyed and pink-eyed rats, cer- tain F! females were back-crossed with a pure pink-eyed male. They produced 46 black-eyed and 39 pink-eyed ; expected, 42 of each. 175 176 INHERITANCE IN RATS. The pink-eyed yellow and red-eyed yellow of rats are complementary loss variations; for when the two varieties are crossed with each other they produce FX offspring which are either gray or black pigmented, according as their yellow parents did or did not transmit the agouti factor. These F! reversionary grays or blacks are paler in pigmenta- tion than ordinary gray or black rats, indicating that neither character in a heterozygous form is the full complement of the other. But it is evident that in homozygous form each is the full complement of the other, since in F 2 and later generations grays and blacks of full intensity are obtained. The F! black-eyed animals (blacks or grays) obtained by crossing pink-eyed yellows with red-eyed yellows, if mated with each other, pro- duce an F 2 generation containing (1) black-eyed young (black or gray), (2) red-eyed yellow young, and (3) pink-eyed yellow young. We have obtained thus far 324 such F 2 young, of which 162 were of class (1), 90 of class (2), and 72 of class (3). If, as suggested, red-eyed yellow and pink-eyed yellow are due to mutually independent Mendelian factors, then F 2 should contain four classes instead of the apparent three; wherefore it seemed probable that one of the three classes was really composite and that the three should be as 9:3:4. On this basis the F 2 expectation would be 182 : 61 : 81 instead of the observed 162 : 72 : 90. Hence there appear to be fewer black-eyed young than are expected. Further, when we came to test the other F 2 classes to discover which of them was com- posite, we found very few individuals which would fall in the hypotheti- cal fourth class transmitting both pink-eyed and red-eyed yellow in the same gamete. Instead of 1 hi 16 as expected, we have been able to discover a much smaller number of double recessives. Both a defi- ciency in double recessives and a deficiency in double dominants (the black-eyed class), which have been observed among the F 2 rats, would be expected if pink-eyed yellow and red-eyed yellow are due to " linked genes," i. e., to factors located near each other in the germ-plasm. For in the cross under consideration each form of yellow enters the F! zygote in a different gamete. Hence, in the gametes arising from such zygotes we should expect the two forms of yellow to show mutual repulsion. If they did so, then the gametes formed by FX zygotes, of the four possible combinations, RP, Rp, rP, and rp, would not be equally numerous, but Rp and rP should be more numerous than RP and rp. That this is true is indicated by the facts presently to be stated. To test the gametic composition of the F 2 yellows, those which were red-eyed were mated with pink-eyed yellows of pure race, and those which were pink-eyed were mated with red-eyed yellows of pure race. For it was known that, since red-eyed yellow is a recessive variation, every red-eyed F 2 yellow must be homozygous for red-eye, but conceivably it might be either heterozygous for pink-eye or might lack it altogether. A cross with the pure pink-eyed yellow race would GAMETIC COUPLING IN YELLOW RATS. 177 decide between these possibilities. Further, it was clear that every pink-eyed F 2 yellow must be homozygous for that character, which is also recessive, but might be either homozygous or heterozygous for red-eye without affecting its appearance, or might even lack the gene for red-eye altogether. A cross with pure red-eyed animals would suffice to show in each case which possibility was realized. In accord- ance with this reasoning the proposed tests have been made hi the case of 45 red-eyed and 40 pink-eyed F 2 yellows. Of the 45 red-eyed yellows tested, 32 have given exclusively black- eyed young (blacks or grays), no test being considered adequate which did not produce 4 or more young; but 13 of the tested animals gave a mixture of black-eyed and of red-eyed young in approximately equal numbers. The former group, numbering 32, evidently lacked the gene for pink-eye, since they always produced atavists in crosses with pink- eyed yellows; the latter group, numbering 13, were evidently hetero- zygous for pink-eye, since only part of their young were atavistic. Of the 40 pink-eyed F 2 yellows which were tested, 27 produced only black-eyed young; these evidently lacked the gene for red-eye. Ten others produced both black-eyed and red-eyed young, being evidently heterozygous for red-eye. Three have produced only red-eyed young, which shows them to be homozygous for red-eye as well as for pink-eye. Hence they are the double recessives, expected to be one-sixteenth of all F 2 rats if no linkage occurs, but less numerous if linkage occurs. We are now in a position to estimate the strength of the linkage shown. If we designate by r the recessive gene for red-eye, and by p the recessive gene for pink-eye, then in the current Mendelian terminol- ogy the following F 2 classes are to be expected in the frequencies shown, if no linkage occurs : Black-eyed. Red-eyed. Pink-eyed. 1 RRPP... 2 RrPP . . . IrrPP 2 rrPp 1 RRpp 2 Rrpp 2 RRPp . . . 1 rrpp 4RrPp 9 3 4 For the present we may pass by the black-eyed classes, since none of these were individually tested. The individual tests already described have shown the existence of the expected two classes of red-eyed and three classes of pink-eyed young, but in proportions very different from those given in the table. Among the red-eyed, instead of the expected 1 rrPP : 2 rrPp, we observe 32 : 13. Among the pink-eyed, where we expect 1 RRpp : 2 Rrpp : 1 rrpp, we observe 27 : 10 : 3. These are very different frequencies from those expected, and they strongly suggest linkage. How strong is the linkage? We may estimate it 178 INHERITANCE IN RATS. from the actual proportions of the four possible kinds of gametes which the FI parents produced. With no linkage these gametes should be of four sorts, all equally numerous, viz, RP + Rp + rP + rp. Linkage would tend to increase the proportion of the two middle classes (Rp and rP, the original combinations) at the expense of the extremes (RP and rp, the double dominant and double recessive classes) . The latter may be called "cross-over" classes, the former " non-cross-over." In producing the 85 F 2 yellow rats which were tested, twice that number of gametes were concerned, viz, 170. From the demonstrated genetic constitution of the tested animals, we can estimate how many cross- over and how many non-cross-over gametes entered into each. Zygotes. Cross-over gametes. Non-cross-over gametes. 32rrPP 13 rrPp 13 64 13 27 RRpp 10 Rrpp 10 54 10 3 rrpp 6 85 29 141 The estimated proportion of cross-over to non-cross-over gametes is seen to be 29 : 141 or 1 : 4.8. In the terminology of Bateson and Pun- nett this would be a reduplication series lying between 1:4:4:1 and 1:5:5:1; in the terminology of Morgan, 17 per cent of the gametes formed by F t individuals are cross-over gametes. We can test this linkage theory in another way. If linkage exists it should modify the proportions of the apparent classes in F 2 as well as of the real classes, which we have just been considering. The apparent classes are three, viz, black-eyed, red-eyed, and pink-eyed, with observed frequencies of 162 : 90 : 72. If no linkage exists the expected frequencies are 182 : 61 : 81, which deviate considerably from the expected frequencies. But if linkage exists, it will lessen the discrep- ancies. Linkage of 17 per cent strength will change the expectations to 164 : 79 : 81 . This alteration shows agreement almost perfect in the case of the black-eyed class, a much reduced discrepancy in the case of the red-eyed class, and no change in the pink-eyed class on the whole a much improved agreement between expected and observed frequencies. Sturtevant has called attention to the fact that double recessives could occur among our F 2 animals only as a result of cross-overs occur- ring simultaneously in the gametes of both parents, a fact which Wright and I considered too obvious to demand comment in our preliminary paper, but recognized in our calculation by counting two cross-over gametes for every double recessive zygote. Sturtevant has questioned the adequacy of our tests in the case of these doubly recessive indi- viduals because apparently he had formed the idea, from studies made on insects, that crossing-over could occur only in the gametogenesis of GAMETIC COUPLING IN YELLOW RATS. 179 one sex. I may say, therefore, that the classification of two animals as double recessives made in our preliminary paper was based on tests which had produced 14 and 9 yellow young respectively. The only possible alternative classification would have involved an expected 1 : 1 ratio of black-eyed to red-eyed young. The chances are overwhelm- ingly great against the observed results being departures due to ran- dom sampling from this expectation. The additional case of a double recessive reported in this paper is so classified on tests which, to the present time, have produced all together 28 yellow young. The num- ber of young produced hi each of the other tests is indicated below. Tests taken to indicate that the parent was of the formula rrPP produced only dark-eyed young (gray or black-coated), as follows: No. of young.... 4 5 6 7 8 9 10 11 12 13 16 17 19 Cases 244522131231 2 = 32 Tests showing the parents to be of the formula rrPp gave the follow- ing numbers in 13 tests: Dark-eyed :Pink-eyed. . .2: 6 4:2 5:4 1:5 3:3 5:3 4:6 1:5 5:5 3:4 3:2 1:5 6:3 Pink-eyed animals were classified as of formula RRpp on the basis of the following tests, which yielded only dark-eyed young: No. of young.... 4 6 7 8 9 10 11 12 13 14 15 16 18 Cases 213222161311 2 = 27 Pink-eyed animals were shown to be of formula Rrpp by the follow- ing tests: Dark-eyed: Red-eyed... 10: 7 8:4 1:4 7:8 6:7 5:6 3:1 4:6 4:2 8:2 Both red-eyed and pink-eyed yellow rats, when crossed with albinos, produce an F\ generation consisting exclusively of black-eyed (black or gray) young. F 2 from the red-eyed cross consisted of black-eyed, red-eyed, and albino young, and F 2 from the pink-eyed cross consisted of black-eyed, pink-eyed, and albino young. If no linkage occurs the expectation in each case is 9 : 3 : 4, and we at first supposed that this was the ratio approximated. But a summary of all litters thus far obtained indicates a probable linkage between albinism and the two yellow variations. Thus, red-eyed non-agouti yellows mated with albinos from our plus- selected hooded race produced 17 black F 1 young. These have given us 58 F 2 young, of which 30 are black-eyed, 18 red-eyed, and 10 albinos. A 9 : 3 : 4 ratio would call for 32.5 : 11 : 14.5. It is evident, therefore, that we have too many red-eyed young and too few black-eyed and albinos. Linkage (in this case, repulsion) between red-eye and albin- ism would tend to increase the number of red-eyed and to decrease the number of black-eyed without changing materially the expecta- tion for albinos; hence, linkage seems probable. Linkage involving 1 cross-over to 3 non-cross-over gametes, or 25 per cent cross-over 180 INHERITANCE IN RATS. gametes, would give an expectation of 29.9 black-eyed : 13.6 red-eyed : 14.5 albinos, which agrees much better with the observed numbers (30 : 18 : 10) than does the 9:3:4 distribution. But if red-eye is linked with albinism as well as with pink-eye, then albinism and pink- eye should be linked with each other. Apparently such is the case, for three F 2 Utters from the cross pink-eye X albino include 12 black, 12 pink-eyed, and 3 albino young. A 9:3:4 ratio (expected if no linkage occurs) would call for 15 black, 5 pink-eyed, and 7 albinos. Linkage of 5 : 1 would call for 14 : 6 : 7, and perfect linkage would call for 14 : 7 : 7. It is evident that the observed numbers of blacks and albinos are too small on any of these hypotheses, but the existence of linkage would tend to diminish the number of blacks and albinos in proportion to the number of pink-eyed, which is the nature of the deviation observed. To determine definitely whether linkage really occurs between the yellow variations and albinism, and if so, what is its strength, further experiments are needed, which are now in progress. It will also be desirable to determine whether the linkage strength is the same in both sexes. SUMMARY. Two yellow variations in rats which have recently arisen as muta- tions show mutual repulsion in heredity. When crossed with each other they produce an F! generation composed exclusively of rever- sionary dark-eyed individuals. The F 2 young are of three apparent classes, dark-eyed, red-eyed, and pink-eyed. Their numerical propor- tions deviate somewhat from the typical 9:3:4 ratio. Further, the proportions of the several expected classes of red-eyed and pink-eyed young do not agree with those usually observed hi an F 2 Mendelian population. But hi both cases the deviations are largely accounted for by the supposition that the genes of the respective yellow variations are "linked" (hi this case showing repulsion) and that the proportion of "cross-over " gametes is about 17 per cent, or mother words, that non- cross-over gametes are about 4.8 times as numerous as cross-over gametes. NOTE. In the foregoing discussion it has been assumed that the ratio of cross-over to non-cross-over gametes is the same among gametes which take part in producing yellows as among those which take part in producing black- eyed individuals. Theoretically it should be slightly different, as the following table will show: Ratio cross-over to non-cross-over gametes. Per cent cross-over gametes. Per cent among gametes pro- ducing yellows. Per cent among gametes produc- ing black-eyed. 1: 1 1:2 1:3 1:4 1:4-4 1:5 1:6 50 33.3 25 20 18.5 16.7 14.3 42.9 29.4 22.6 18.4 17. 1 15.5 13.4 55.6 36.8 27.3 21.6 19. 8 -> 17.8 15.2 TABLES. 181 TABLES. Table 141 shows the classification of extracted hooded first crossing hooded rats of the plus-selected series with wild rats. TABLE 141. young obtained from Hooded grandparents. Grade of hooded grandchildren. Total hooded. Total non- hooded. Means of hooded. 11 u 2 21 21 2f 8 3* 3} :*! 4 95513, +41, gen. 10 1 3 1 2 1 1 7 2 8 4 6 3 1 5 4 1 1 7 6 3 1 1 41 22 5 3 2 107 68 27 12 5 3.05 3.28 3.51 3.17 3.37 o"6348, +4, gen. 10 9 6955, +4, gen. 12 9 5513, +4J, and 9 6600, +41, gen. 12 3 95513, -Hi, and 96955, +4, gen. 12 1 1 Totals .... 1 4 2 2 9 14 11 12 16 2 73 219 3.17 Table 142 shows the classification of extracted hooded second 2 young obtained from crossing first F 2 hooded rats (table 141) with wild rats. The hooded grandparents were themselves grandchildren of 9 5513, +4|, generation 10, on the side of both parents. TABLE 142. Hooded grand- parents. Grade of hooded grandchildren. Total hooded. Total non- hooded. Means of hooded. 2 2J 2i 2f 3 3i 3* 31 4 99619, +2.. 1 1 2 1 2 5 1 1 2 2 4 11 6 2 i 3 4 7 8 7 30 1 4 1 5 2 5 13 10 30 22 16 8 28 24 22 104 68 42 3.37 3.40 3.06 3.62 3.47 3.55 3.70 0*9686, +2J. . 1 1 99620, +2i... 99729, +21... 1 1 2 1 c?9727, +3.... 9 9728, +3 .... 99621, +31... 1 i 2 1 3 1 Totals . . . 1 2 8 4 6 13 2S 1! 98 296 3.47 Table 143 shows the classification of extracted hooded second F 2 young obtained from crossing first Fa hooded rats (table 141) with wild rats. The hooded grandparents were themselves grandchildren of cf 6348, +4, generation 10, on the side of both parents. TABLE 143. Hooded grand- parents. Grade of hooded grand- children. Total hooded. Total non- hooded. Means of hooded. If 1 2 2 1 21 1 2J 2! 3 3 6 l 3} 4 34 15 4 si 6 4 cf9639, +2. ... 1 39 6 1 27 16 21 9 2 110 16 10 76 47 74 40 3 3.24 3.17 3.50 2.90 3.28 3.48 3.36 3.87 99704, +2i. . . . 9 9765, +3 . . 1 99747, +3i 9 9703, +3i 1 7 i i i 1 2 1 1 7 2 5 2 1 1 1 1 4 6 4 2 4 2